ポアソン回帰モデルの推定
- まず、全変数を投入したモデル
npub.poisson1
を推定します。
npub.poisson1 <- glm(art~fem+mar+kid5+phd+ment, data=npub,
family = poisson(link = "log"))
summary(npub.poisson1)
Call:
glm(formula = art ~ fem + mar + kid5 + phd + ment, family = poisson(link = "log"),
data = npub)
Deviance Residuals:
Min 1Q Median 3Q Max
-3.5672 -1.5398 -0.3660 0.5722 5.4467
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.304617 0.102981 2.958 0.0031 **
femWomen -0.224594 0.054613 -4.112 3.92e-05 ***
marMarried 0.155243 0.061374 2.529 0.0114 *
kid5 -0.184883 0.040127 -4.607 4.08e-06 ***
phd 0.012823 0.026397 0.486 0.6271
ment 0.025543 0.002006 12.733 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for poisson family taken to be 1)
Null deviance: 1817.4 on 914 degrees of freedom
Residual deviance: 1634.4 on 909 degrees of freedom
AIC: 3314.1
Number of Fisher Scoring iterations: 5
- 変数
ment
とart
の散布図にモデル推定結果を重ねます。
ggplot(npub, aes(x = ment, y = art, col = 2)) + geom_point() +
geom_smooth(method = "glm", method.args = list(family = "poisson")) +
guides(colour = "none")
![](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAkAAAAJACAYAAABlmtk2AAAEGWlDQ1BrQ0dDb2xvclNwYWNlR2VuZXJpY1JHQgAAOI2NVV1oHFUUPrtzZyMkzlNsNIV0qD8NJQ2TVjShtLp/3d02bpZJNtoi6GT27s6Yyc44M7v9oU9FUHwx6psUxL+3gCAo9Q/bPrQvlQol2tQgKD60+INQ6Ium65k7M5lpurHeZe58853vnnvuuWfvBei5qliWkRQBFpquLRcy4nOHj4g9K5CEh6AXBqFXUR0rXalMAjZPC3e1W99Dwntf2dXd/p+tt0YdFSBxH2Kz5qgLiI8B8KdVy3YBevqRHz/qWh72Yui3MUDEL3q44WPXw3M+fo1pZuQs4tOIBVVTaoiXEI/MxfhGDPsxsNZfoE1q66ro5aJim3XdoLFw72H+n23BaIXzbcOnz5mfPoTvYVz7KzUl5+FRxEuqkp9G/Ajia219thzg25abkRE/BpDc3pqvphHvRFys2weqvp+krbWKIX7nhDbzLOItiM8358pTwdirqpPFnMF2xLc1WvLyOwTAibpbmvHHcvttU57y5+XqNZrLe3lE/Pq8eUj2fXKfOe3pfOjzhJYtB/yll5SDFcSDiH+hRkH25+L+sdxKEAMZahrlSX8ukqMOWy/jXW2m6M9LDBc31B9LFuv6gVKg/0Szi3KAr1kGq1GMjU/aLbnq6/lRxc4XfJ98hTargX++DbMJBSiYMIe9Ck1YAxFkKEAG3xbYaKmDDgYyFK0UGYpfoWYXG+fAPPI6tJnNwb7ClP7IyF+D+bjOtCpkhz6CFrIa/I6sFtNl8auFXGMTP34sNwI/JhkgEtmDz14ySfaRcTIBInmKPE32kxyyE2Tv+thKbEVePDfW/byMM1Kmm0XdObS7oGD/MypMXFPXrCwOtoYjyyn7BV29/MZfsVzpLDdRtuIZnbpXzvlf+ev8MvYr/Gqk4H/kV/G3csdazLuyTMPsbFhzd1UabQbjFvDRmcWJxR3zcfHkVw9GfpbJmeev9F08WW8uDkaslwX6avlWGU6NRKz0g/SHtCy9J30o/ca9zX3Kfc19zn3BXQKRO8ud477hLnAfc1/G9mrzGlrfexZ5GLdn6ZZrrEohI2wVHhZywjbhUWEy8icMCGNCUdiBlq3r+xafL549HQ5jH+an+1y+LlYBifuxAvRN/lVVVOlwlCkdVm9NOL5BE4wkQ2SMlDZU97hX86EilU/lUmkQUztTE6mx1EEPh7OmdqBtAvv8HdWpbrJS6tJj3n0CWdM6busNzRV3S9KTYhqvNiqWmuroiKgYhshMjmhTh9ptWhsF7970j/SbMrsPE1suR5z7DMC+P/Hs+y7ijrQAlhyAgccjbhjPygfeBTjzhNqy28EdkUh8C+DU9+z2v/oyeH791OncxHOs5y2AtTc7nb/f73TWPkD/qwBnjX8BoJ98VQNcC+8AAEAASURBVHgB7J0HnFTV2f8f2N4bC+wuLL0X6SAiVRF71+jfGjVFYyxpvjFqXjUaTUxiyRtfNcmbRE0Ue1eaiCAgvUmHpSws29hegf/5XbzL7OzcmTv9zp3fw2eZmXPPPeV77tz7zDnPeZ5OJ5QIhQRIgARIgARIgASiiEDnKOoru0oCJEACJEACJEACGgEqQLwQSIAESIAESIAEoo4AFaCoG3J2mARIgARIgARIgAoQrwESIAESIAESIIGoI0AFKOqGnB0mARIgARIgARKgAsRrgARIgARIgARIIOoIUAGKuiFnh0mABEiABEiABKgA8RogARIgARIgARKIOgJUgKJuyNlhEiABEiABEiABKkC8BkiABEiABEiABKKOABWgqBtydpgESIAESIAESIAKEK8BEiABEiABEiCBqCNABSjqhpwdJgESIAESIAESoALEa4AESIAESIAESCDqCMRGXY9D3OHKykppbGwMWK3p6emSkpKilVdeXi7Nzc0BK9sqBeXl5UlVVZXU19dbpUkBa0fXrl0lJiZGTpw4IYcPHw5YuVYpKCEhQbKzs6WkpESOHz9ulWYFpB2xsbGSm5urldXQ0CBHjx4NSLlWKiQjI0Pi4uKkrKzMSs0KSFtSU1MlLS1NKwtjhzG0m3Tr1k3q6uqktrbWbl2TnJwciY+P1/qFeyfuoRA8L3wVzgD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EqAD5So7nkQAJkAAJkAAJRCwBKkARO3RsOAmQAAmQAAmQgK8EYn09keeRAAmQAAl0JHCksk4+/Gq7rN91WGrrmyU7PUnGDc6X8yYNlNSk+I4nMIUESCAsBKgAhQU7KyUBErAjgfeWbpU/v71SWlqPt3Vv96FKWbWtWF6Zt0F+dcM0GT+4oO0Y35AACYSPAJfAwseeNZMACdiIwEfLt8uf5i5vp/w4dq9GzQbd/+IC2bi7xDGZ70mABMJEgApQmMCzWhIgAfsQqKxpkOfeWumxQ63Hjsvv/7NUjh0/NUPk8SRmIAESCAoBKkBBwcpCSYAEoonA/NW7pbG51VSX9x+plnU7DpvKy0wkQALBI0AFKHhsWTIJkECUEPhmb6lXPd3iZX6vCmdmEiABUwSoAJnCxEwkQAIkYEygrrHF+KCLI/VN3uV3UQSTSIAE/CRABchPgDydBEiABLpmpXgFITcz2av8zEwCJBB4AlSAAs+UJZIACUQZgQlDvNvaPmFIjygjxO6SgPUIUAGy3piwRSRAAhFG4IzhhdIvP8tUq2eN6SM9ctNN5WUmEiCB4BGgAhQ8tiyZBEggSgh07txJHrp5hmSkJrrtcZ+8TLnrytPd5uFBEiCB0BCgAhQazqyFBEjA5gQwq/P8Ty6QiUM7Lm917tRJhcIYIM/edT7DYdj8OmD3IocAQ2FEzlixpSRAAhYn0C0rVR7/3llysKxaNu4qkZoGFQssLUlGD8zTXi3efDaPBKKKABWgqBpudpYESCAUBAq6pAv+KCRAAtYlwCUw644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkAlSAggSWxZIACZAACZAACViXABUg644NW0YCJEACJEACJBAkArFBKpfFOhCIiYlx+OTf206dOrUV0LlzZwlk2W0FW+AN+mnXvul47dg/XJMQ9M3xWtX7HMmvet/QB7ten/qY2fHa1PuG8bPzvdOufXMcP1yfJ06cwFD6JZ1UIf6X4lcT7H1yfX29JCcn27uT7B0JkAAJkAAJRBgBKkBBHrCqqio5duxYwGqBMpWYmKiVV11dLa2trQEr2yoFZWdnS11dnTQ1NVmlSQFrR0ZGhjY7gt8dlZWVASvXKgXFxcVJWlqa1je7/bbCr06MH6S5uVlqa2utgj1g7UhJSdGuT9xb7Ca4b+o/RjF2GEO7SWZmpnbfbGhosFvXJD09XWJjTy5a4d6p31/wvPBVuATmKzmT5x0/fjygD/KEhIS2mltaWmz5JUYHodjZUQFqGzz1xs79w8MF176dRL/5ok/4UWPH8YOSgCUUO/YNyrkudr6/2LVvjvcTXJ+6AqSPqS+vNIL2hRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCVIAievjYeBIgARIgARIgAV8IUAHyhRrPIQESIAESIAESiGgCsRHdejaeBGxAoKq2UfYcqpSW1uOSn5smBV3SbdArdoEESIAErE2ACpC1x4etszGBfSVV8vy7X8uKbw7IiROnOtovP0tuu3CsTBjS41Qi35EACZAACQSUAJfAAoqThZGAOQJfb9knP3zqfVm+pb3yg7N3FVfKff87X15ftMlcYcxFAiRAAiTgNQEqQF4j4wkk4B+Biup6ufOpt6ShudVtQc+/u0pWbSt2m4cHSYAESIAEfCNABcg3bjyLBHwm8Nf3lkt1XaOp81/6YLWpfMxEAiRAAiTgHQEqQN7xYm4S8JvA/JXbTZexfX+5HKmsM52fGUmABEiABMwRoAJkjhNzkUBACBw/fkIOHDnqVVmHymu8ys/MJEACJEACnglQAfLMiDlIIGAEOnfuJHGxMV6VF+9lfq8KZ2YSIAESiFICVICidODZ7fARGNK7m+nKY2M6S6/umabzMyMJkAAJkIA5AlSAzHFiLhIIGIFLZ4w0XdbU03pJcmKc6fzMSAIkQAIkYI4AFSBznJiLBAJG4LLpI2XUwAKP5aUlx2sOET1mZAYSIAESIAGvCVAB8hoZTyAB/whgWevPP71chvXpalhQdnqSPPmD2dItK9UwDw+QAAmQAAn4ToChMHxnxzNJwGcCWenJ8qc758hnX++ST1fulF0HK6T12HEtDhiWvS6bNlRSk+J9Lp8nkgAJkAAJuCdABcg9Hx4lgaARiOncWc6dOED7C1olLJgESIAESMAlAS6BucTCRBIgARIgARIgATsToAJk59Fl30iABEiABEiABFwSoALkEgsTSYAESIAESIAE7EyACpCdR5d9IwESIAESIAEScEmACpBLLEwkARIgARIgARKwMwEqQHYeXfaNBEiABEiABEjAJQEqQC6xMJEESIAESIAESMDOBKgA2Xl02TcSIAESIAESIAGXBGztCLG+vl5++tOfSvfu3eXBBx90CQB5XnzxRZfH9MTzzz9f+vfvr3/kKwmQAAmQAAmQQIQTsK0CdOLECXn44Ydl48aNEhtr3M2dO3fKG2+84XYYR40aRQXILSEeJAESIAESIIHIImCsGURWP9q1tqGhQZ5++mlZunRpu3RXH3bs2KEljx07VmbOnOkqiwwYMMBlOhNJgARIgARIgAQik4DtFKBVq1bJE088IYcPH5bOKtbS8ePH3Y6MrgBB+bnooovc5uVBEiABEiABEiABexCwlRH0vHnz5J577tGUn4kTJ8q9997rcZSwBAYZNGiQx7zMQAIkQAIkQAIkYA8CtpoBqqiokPz8fLnhhhvkvPPOkyVLlrgdpdbWVtm9e7dmI9S3b18tb3V1tSQkJGh/bk/mQRIgARIgARIggYglYCsFaNasWXL55Ze7NXp2HKmioiJpaWmRwsJCefXVV+Xtt9+W8vJybekMaddff73Mnj3b8RSX72+77TZpbm52eezmm2+W008/3eUxXxIdDbrT0tIExt52lJSUFElMTLRd17Asq0t2drb+1javev+ysrJsd2126tSpbZzi4+PFjuOH+wv6ace+xcTEtI0f7i9JSUltn+3yBmOHfuH6tJs4PvtwfwmE2EoB6tKli1dMdPufffv2yUsvvSQ4H8bQe/fu1f4eeeQRWbNmjdx3331uy12xYoU0NTW5zHPJJZcEbTbJjhe5DhEXu+MFr6fb5RU3Ksw02lXsfG1izPAwdXyg2m0c7XxtYqzi4uLsNmRt/bH7vRMdDdT1aSsFqO0KMPlGt//BTMpjjz0m2O4OwawKZoP+9Kc/yYcffiiTJk2S6dOna8f4HwmQAAmQAAmQQOQTiGoF6MYbbxQsm2VkZGi2Q/pw4tf5ZZddps0CQRF6+eWX3SpAML42EuxCKykpMTrsdTqUteTkZO28yspKw6U3rwu20AndunUT2GLBnYHdBLOMmDmAkn3kyBG7dU+besf0dGlpqccdmJHWefyyzsnJ0Zrd2NgoVVVVkdYFj+1NT0/XZl5hT2k3wbJXamqq1i2MHcbQbpKbmytw7ltXV2e3rmnLsvrMHe6duvkHnhe+SlQrQFAmhgwZYsjurLPO0maC9uzZo93MdfsG5xPcDQCUlEB+0fRBRxugXHna5u/c1kj5jH7atW/6GNixf/r1acdr03G87Hp9Oo6ffp3a5VXvG/pj1/Gzc98cxw/fRcfPvl6jpywyfS3Bxufpig0MnO2oUdt46Ng1EiABEiABEnBLIKpngObOnavt+pozZ4707t27Ayh96QrTwpgtopAACZAACZAACdiDQFTPAM2fP19eeeUV+cc//uFyNL/88kstfdiwYS6PM5EESIAESIAESCAyCUS1AjRjxgxt1BYuXCi7du1qN4KrV68WzBBBbrnllnbH+IEESIAESIAESCCyCUT1EtiVV14py5Ytk7Vr18pNN92k+QCCHyBsj1+0aJE2srfffjvDZET2Nc7WkwAJkAAJkEAHAlGtAGE78pNPPqktg8ETNGZ98Afp2bOn/PjHP9Z8AHWgxgQSIAESIAESIIGIJmBrBWjq1Kke44Eh3AKWuOATqLi4WI4ePSqIC6b7i4jo0WXjSYAESIAESIAEXBKwtQLksscGiXByhvhf+KOQAAmQAAmQAAnYm0BUG0Hbe2jZOxIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwJUgIzIMJ0ESIAESIAESMC2BKgA2XZo2TESIAESIAESIAEjAlSAjMgwnQRIgARIgARIwLYEqADZdmjZMRIgARIgARIgASMCVICMyDCdBEiABEiABEjAtgSoANl2aNkxEiABEiABEiABIwKdTigxOsh0/wnU1dVJUlKS/wV9W0KnTp0Ef5Djx49/m2qvl86dO9u2b/r44Wtn16+enccPfYPYdfz0e4udr02Mn13vnXYeP/3e6Tx++ncS6d5KrLcnML93BI4dOyZQggIlCQkJEh8frxXX2NgoKN9ukpaWJs3NzdLS0mK3rklKSkqbAhvI68IqoGJjYzWFv76+3nYKHm60GD9Ia2ur4PtnN0lMTBT0E+NnN8F9E/dPSFNTkzaGdusjrk/cN3H/tJskJydLTEyM1i3HeyeeF74KFSBfyZk8DwpKIG+UuDnpClBDQ4MtL3Rc0LhB2fEmjC+xLrW1tfpb27ziAYMZT9yg7PYrG8qdowJkx/HDAyYuLk7s2LfU1NR2ChDun3YTXJ9Qfuw4fri3OCpA+iylPwoQbYDs9g1gf0iABEiABEiABDwSoALkEREzkAAJkAAJkAAJ+EPg6NGj/pwelHOpAAUFKwslARIgARIgARIAAZiBVFRUWA4GbYAsNyTWaFBtfZO8vWSrLNu0T0oq6yQhPlYG9cyR808fKOMHF1ijkWxF1BLYuLtE3lu6TbbsPSL1jS2Sm5kiE4f2kMumDpGstMDtuoxawOw4CQSIAGwBDx06FKDSAlsMFaDA8rRFaZvUw+WBvy2Sqtr2u1xKKmrli/VFMnNMH/n5NWdIfBwvH1sMeAR14pi6mT775gpN+XFsdlVdk+w8WKGU9m/kVzdMk0lKGaKQAAmEn0B5ebm2Mw2bCKwmXAKz2oiEuT1Fh4/KL56f10H5cWzWwjV75Hf/XuqYxPckEBICz7+7qoPy41gxZoMe/OtCbWbIMZ3vSYAEQk8AO+0qKytDX7HJGqkAmQQVLdmefWuFNDS3euzuAqUErd5W7DEfM5BAoAjsPFAuby7e4rG41mPH5Y+vf+UxHzOQAAkEjwCWvg4fPhy8CgJQMhWgAEC0SxEllbWyZrv5tdqPlu+wS9fZjwgg8PHKnaZbuau4UrbtKzOdnxlJgAQCS6CsrMzyzmypAAV2zCO6tO37y71q//b9fMB4BYyZ/SLg/fXp3fXsV+N4MgmQQBsBOLG14rb3tgZ++4YKkDORKP7cZGLpyxFPU4v9wnA49o/vrUXA++vT81KutXrI1pBA5BPA0ldJSUlEdIQKUEQMU2ga2T3Hu5gq3bNTQ9Mw1kICikBejnfXG69PXjYkEHoCR44csfzSl06FCpBOgq8ypFcXyUxNNE1i8vCepvMyIwn4S2Dy8ELTRSTExciYgfmm8zMjCZCA/wRqamqkurra/4JCVAIVoBCBjoRqYlSg1ZvPHW2qqTnpSXLxlMGm8jITCQSCwKyxfaWwa4apoq45a4QkJ8aZystMJEAC/hNobW0VzP5EklABiqTRCkFbLzxjkObt2V1VqUnx8sitsyQpgQ8Yd5x4LLAEYmM6y8O3zJRsD56ezxxZKNedfVpgK2dpJEACbgnA7ufYsciyC6UC5HZIo/PgT66eLPdedbrLBw087P7lJxfI4MIu0QmHvQ4rgcJuGfL8Ty+U6aN7S+dOndq1JS05Xn5w8Th56KYZ0rlz+2PtMvIDCZBAQAlUVVVJXV1dQMsMRWHW800dil6zDo8ELpg8SM6bNFC2qa3uh1UIjCQVC2xAjxzJyUj2eC4zkEAwCXRR1+CDN06XqssbZavy9aPFAstKkSFKKY9Rs0QUEiCB0BFoaWmR0tLS0FUYwJqoAAUQpt2Kwq/oIb1ytT+79Y39iXwCGcpgHwFQKSRAAuEhcOLECS3QKba+R6Lw51IkjhrbTAIkQAIkQAJhJlBRUSGNje2DZoe5SV5VTwXIK1zMTAIkQAIkQAIkgECnUIAiWagARfLose0kQAIkQAIkEGIC2O2FQKdYAotkoQIUyaPHtpMACZAACZBAiAlEkrdnd2ioALmjw2MkQAIkQAIkQAJtBODpGR6f7SBUgOwwiuwDCZAACZAACQSZALa8R5q3Z3dIqAC5o8NjJEACJEACJEACmr3PoUOHJFK3vLsaQipArqgwjQRIgARIgARIoI1ApG95b+uIwxsqQA4w+NZ/ArUNzdLQ1OJ/QSyBBEiABEjAEgTq6+ulvLzcEm0JZCPoCTqQNKO0rIrqenll3kZZtHaPHK096RSrmwpNcNa4fnL1zOGC4KkUEiABEiCByCOALe9Y+rKjUAGy46iGsE+bdpfI/S8tkJr65na1llTWKaVog8xftUue+MHZUtgts91xfiABEiABErA+ASg/kRbl3SxVLoGZJcV8HQggSOovX+yo/DhmhCL0i+fnCZbGKCRAAiRAApFDoLKyUrD8ZVehAmTXkQ1Bv/724RpTig2UoNcXbQpBi1gFCZAACZBAIAgg1EWkRnk3238qQGZJMV87As0trbJkQ1G7NHcf5q/a7e4wj5EACZAACViEgB7qwiLNCVozqAAFDa29Cy4ur5WmlmOmO4nlssbmVtP5mZEESIAESCA8BEpKSgROD+0uVIDsPsJB6p8vQfB8OSdIzWexJEACJEACLggcPXpUamtrXRyxXxIVIPuNaUh61D07VWJjzF8+OelJkpQQF5K2sRISIAESIAHvCTQ2Ntre7seRivknmONZfB/1BKDMTBzawzSH6aP7mM7LjCRAAiRAAqEloPv7iaaZeipAob3GbFXbd88bLfGxMR77lJGSINecNcJjPmYgARIgARIID4FosftxpEsFyJEG33tFoE9eljxw4zS3SlBacrw8eussyU5L8qpsZiYBEiABEggNAcT5iha7H0eiVIAcafC91wTOGFEoz//0QpmiXh1tghLiYmT2+H7y4s8ulmF9unpdLk8gARIgARIIPgH4+ykrKwt+RRasgaEwLDgokdak3t0z5eFbZmpBUOH0sHOnTtI9J9XtzFCk9ZHtJQESIAG7EdDtfuzWL7P9oQJklhTzeSQAw2goQxQSIAESIAHrE0Ccr9bW6PXPxiUw61+jbCEJkAAJkAAJBJRAeXm5reN8mYFFBcgMJeYhARIgARIgAZsQqKurEyhA0S5UgKL9CmD/SYAESIAEooZAc3OzYOmLIkIFiFcBCZAACZAACUQBgePHj2vKD14pVIB4DZAACZAACZBAVBCAs8Ompqao6KuZTnIGyAwl5iEBEiABEiCBCCaAIKc1NTUR3IPAN50KUOCZskQSIAESIAESsAwBODs8cuSIZdpjlYbY2g9QfX29/PSnP5Xu3bvLgw8+aMgcU4JvvPGGrFq1SiorK2XAgAEyatQomTNnjsTEeI51ZVgwD5AACZAACZBAGAnAzw+Nnl0PgG0VIES0ffjhh2Xjxo0SG2vcTUwL3n777bJ//36NUHZ2tnzyySfa37Jly+Shhx6S+Ph41/SYGvUEjiljwn0lVVLX0Cw5GcmSl5MW9UwIgARIwBoE8BwsLi6OameH7kbCWDNwd5bFj2G67+mnn5alS5d6bOkjjzyiKT8TJ06UBx54QDIyMuTgwYNy//33yxdffCHPPPOMNovksSBmiCoCjc2t8sq8DfLe0m1SU3/KqLCgS5r8v7NHypyJA6KKBztLAiRgPQIwem5sbLRewyzSooDbAL3wwgty9913y5YtW0x3EUrI5MmT5fHHHzd9jlFGLGPdcMMN8uGHH0rnzu67hzauXLlSkpKS5NFHH9WUH5RbUFAgf/jDH7Tlr48//piGY0awozS9qrZR7nz6I00BclR+gONgWY08+e+l8ttXlgh+fVFIgARIIBwEsLpRXV0djqojpk73GoIP3Xj33Xe12Zc9e/aYPvvzzz+Xr776yiulyVXh8+bNk3vuuUcOHz4smNG59957XWVrS0O9kGnTpkliYqL2Xv8PS2ETJkwQOI2CEkQhAZ3Ao/9cLLsOVugfXb5+9vUueXX+RpfHmEgCJEACwSQA+1caPXsmHHAFyHOVp3IgEu3WrVtl/fr1WmJycvKpgz68q6iokPz8fLnvvvvkd7/7nWRlZbktZfPmzdpxKEuuBAoQZMOGDa4OMy0KCazaViyrt5vzovryZ+ul1mF5LApxscskQAIhJkBPz+aB+2UDdP7558uCBQva1dbS0qJ9vvTSSz0uQSGvo0fKcePGtSvL2w+zZs2Syy+/3K3Rs2OZsPWBZGa6jmCup+sG0o7nOr6HnZDeb8d0vJ8xY4YMHDjQOdnnz3FxcW3nQmFMSEho+2ynN+iXFXfgfbX55DVjhnVTyzHZsLdCzpl4avw7derUdmpamv0MpvUxS01Ntd0SoOOSOr6Hdhw/9Av9tGvf9C8fZvzdbY7R80XaK5bdsQKCvlmtf9hM5M91pd9bMCa4vwRC/FKAnnrqKRk5cqTLh7+RQmDU6OHDh8sll1xidNhUepcuXUzl0zMhIBxEV3T0dP01PT1de6vn09OdX1966SVD75p9+/aVMWPGOJ8SkM+wXbKrOC9JWqWfB8u9cyRWcrTe5ZcVilCgvsRWYePYjpSUFMePtnuPh4udx8/OfcPFaNX7i79flL1792rPYyv2D8p1oK6rQJXjlwI0ePBg+ctf/iJff/1127jBXmbfvn1y7rnnSmFhYVu6qzcAghtlnz595KqrrvK4ZOWqDF/TMPOkW8cbaaU6ZLoO95Wy/c6L89IvVKyX+e1HjD0iARIIBQH4+oHhM8U8Ab8UIFRzyy23aH96lVgWgwJ0xx13CN5bVTDNixkUbJk3UnD0dE9+gP74xz+2W8pz7HPv3r0FtkmBEix76do9LPzh5MpuAgN0zLrp/K3Uv8KuabLUiwblZye1G3+4WcBULqaq4XTTbqIvDeFG7Li8bYd+YtwwfhDYWdTW1tqhW+36gB+k6Kcddw/hvqnbmWLsMIZ2EYwXTDrwTMMzwdsVmFBwwL3Bn2chVmT0ZT3cO/Vdtnhe+Cp+K0DOFV999dUyadIkGTRokPMhy33Gkhnse4zio+jpnqbzYXtkJBgofabJKI836Y42P7jI7fQlduSAL7EVFaAZo3vLv+ebM4rPSU+SEX26GPbDiv1zHAN/3qNvdlOA9JsvuGADhx3HD0oCfhzasW94AOti1fuL3j5vXvF8wXNMVwjwvbPqD2N/rivH+wnK0fvrDSvnvAHfBYadVfAF9Ne//lWwFc/KotsM6YqOc1v1X0GedpM5n8fP9iXQvyBbLjrDnHJ/x2UTJT6WoVTsezWwZyQQXgJQdODpORDKQHh7Ep7aA64Avfzyy3LgwAF58cUXLb9DqWvXrhr13bt3u6Svpw8ZMsTlcSZGJ4EfKcVm1pg+hp2HgfOPLpsg00f1NszDAyRAAiTgDwEoPQxz4Q9BkYAqQJgWLi0t1Vo0ZcoUS25jdsSlL13Nnz/fMVl7j+m2hQsXau8RGJVCAjqB2JjOcv8N0+Th786QMQPz2mZ50pMTNMXohZ9dKJdNHapn5ysJkAAJBJwAjJ4DaV4R8AZGQIEBtQGC8dzUqVM130AIMQElwtF3htV4wFYJRso7duzQvD1j55our7zyipSXl0uvXr00r9J6Ol9JQCcwZWQvwR+kufVYmyKkH+crCZAACQSDQFlZWcQY4X+29qD07poqQ3vlBAOFX2UGdAYILUFMLVhrQzu99dZbLb2bAEsVt912mzZT9dhjj8mDDz4o//rXv+TnP/+5ZscEozm8d3Re5xdtnmxbArT1se3QsmMkYCkCsE31ZzdVKDsD5eetr4rk2Q+/kW/2W2/Xa0BngAAWTgWff/55TXH4+9//Lm+//bbAhmbAgAHabIu7LeWYPTrzzDNDOT7ajBW2sUMBWrRokfaHBmBmCHHF4OiRQgIkQAIkQALhJoCNRYjwHgkyf12xpvygrS2tx+WZ9zZL9+55MnZQvmWaH3AF6Cc/+Yl89NFHbR2EPxAEOsWfJ/n1r38dUAUICtWSJUs8VSujR4+WuXPnakte2E4I4+ju3btbevnOY6eYgQRIgARIwDYE9BhfkbDja8H6Ynlj2d527FuOHZdfvrhAHr1lpowfUtDuWLg+BFwBCldHAlFvTk6O4I9CAiRAAiRAAlYhgA1GcHSIV6vL5xsPydyle102s0XZSiKgtG0VIGyD99XZkR56wiU5JpIACZAACZBAlBHAZiJsd7eid2fnoVi86bD8Z8ke5+S2z/Ch9sNLxrd9DvebgM8A0WlguIeU9ZMACZAACdiFAKK7I2ST1QUzP+6Un2nD8+SuKyZZqhsB3wXmT+8iYW3Tn/7xXBIgARIgARIwSwB+9SIh5tyiDe6VnzOGdJXrZvS33I7qgM8AOQ7s3r17BdorjLcc43hA0YELb6xnIuglrNo//PBDmTx5stx///2ORfA9CZAACZAACUQdAWwgioSAyTB4NrL5waCdPihXrpvez3LKD9oWFAVo+fLlct9998nixYtRh2kZP946a4OmG82MJEACJEACJBBAApj1OXLkSABLDE5R2OruvNvLsaZJSvm5fqb1Zn70NgZcAYLGeumll2ozP3olZl7z8/Nl8ODBZrIyDwlELYHDFbXyTVGpNDS1SG5miozs200S4gP+NY5avpHa8dKjdbJ5b6nUNzZLl4xkGdmvuyTyuojI4UR4CzgStrroTg6N2gnl5wal/HRWDoetKgG/c8KhIJa9IIi1ddFFF0lSUpJ873vf04KjvvTSS9qyV1FRkbz++uuya9cu6du3r2zdulXgeZlCAiTQkcDBsmp55o0V8vXWg+0OJiXEyndmjpBrzx4hMZ0tZdLXrp38EBwCUIife2uFLNu0v10FCXExctWM4XL9OacJYtdRIoMAzEUQTNzq9rAfrTog763cZwh18mDY/PSztPKDxgdcAVq9erUGZfbs2fLpp5+2AXriiSc0ZWfgwIEyYcIELR1hJubMmSMrVqyQp556Sls2azuBb0iABDQCOw6Uy0/+/KnUNjR3INLQ1Cp//3itNiv0yK0zqQR1IGTfhL2Hj8o9z34sVXVNHTrZ1HJM/vXZejUrdEQe/95ZEhcb0yEPE6xFAHax8PXjaC9rrRaebM27K4rk49Xtf4g5tlMzeLaozY9jO/E+4D8NEFgUctddd2mv+n+nn3669hbhJnRB2AxEYu/Xr588/PDDsmePsf8A/Ry+kkA0EWhuaZUH/7rQpfLjyGH5lgPyr0/XOybxvY0JtCqvurguXCk/jt1es/2Q/O2jtY5JfG9BArqjQ6v7+pm7dI9b5efMod0sa/DsatgDqgBh8KDBQhD7y1EGDRqkfdywYYNjssD5IaKww88B4oZRSIAEThH4eMVOKamsO5Xg5t1rCzdptkFusvCQTQgsWL1bDpRWm+rNm4u3SLWLWSJTJzNT0AlguQs2P746EA56A1UFaOOri3fJgvXGtknThneXa6f1teRuLyNGAVWAYMOjh5KIjW2/umakAKFh06ZN09q3ceNGo3YynQSiksDyze1tO9xBwLLH2h0n7e/c5eOxyCfwlRfXBWaLVqvwAxRrEoDNLIKcWlWOK+Xnnwt3yhebjYOwzhyZJ9dMjSzlB7wDqgChQH0nl/NyFiLCQ2DsDEMvR0lOTtY+bt682TGZ70kg6gmYnf3RQZVU1upv+WpjAiXK+Nkb4XXhDa3Q5cVW95qamtBV6GVNx5Ty/Nd52+WrbaWGZ54zukCumtLH8LiVDwRNAXr++efb9RszQJgVgqHXF1980e7YBx98oH1OS0trl84PJBDtBBLVLi9vJIlbn73BFbF5kxLivGq7t/m9KpyZfSJQXl4ucHZoVWlpPS5/+WSbrN5ZbtjEC8b3lEtP72V43OoHAq4AXX/99doa4Ny5c+WKK66QVatWaQywPHbGGWdo72+//XYtuBvWFd9//3154403tPT+/ftbnRfbRwIhJTCkMNer+gYVdvEqPzNHJoHBvbwbZ14X1hpn+MuDAmRVaWw+Js98sEU2FVUaNvEypfhAAYpkCbgCNHXqVLnzzjs1Jm+++abmB0gHdO+992pvsVOsZ8+ekpeXpx1HvBMIlCcKCZDAKQLnThogZt2IDeudK33ysk6dzHe2JTBnwgDp3NncldG/IFsG9cyxLYtI61h1dbXozzwrtr22sUX++O4m2VFsbGR/9Zl9ZLZa+op0CbgCBCCPP/64/OhHP9J2eGGLuy4XXnih3HHHHdpH+DpADDBdbrvtNpkyZYr+ka8kQAKKAB5eV80c7pEFHN/dfdVJVxMeMzNDxBMo7JYh15090mM/4mI7y71XT46onTkeOxXBGRDiQncUbMVuHK1rlqfe3iRFpa53nkLlvl45OJwxIs+Kzfe6Td4ZGJgsHkbNzz77rPzmN7+RdevWtZ3VSbnEfu6552Ts2LHyzjvvyNq1awXLXtdee63ceuutbfn4hgRI4BSB2y4Yq5yjnZC5n7veJJCZmii/vnmG9MvPPnUS39mewI1zRskxdV28Om+DnHDR27TkeHnopukymMuiLuiEPgmuXqwc4qK0qlGefn+zlFV3dKwJWphxvOWsATK2v3fLr6Enbb7GoChAevXp6emCJTFnufnmmwV/FBIgAc8EcOP54SXjZdbYPvLBsu2yBTGfVCywrlkpMmloD7nwjEGSkhjvuSDmsBUB/KC85fwxMn10b3l/6TYtFlid8haem5ksE4b0kIvUdZGWnGCrPkdqZxDfCz7yrBriYn9ZnTzz/hapaWhxiThOhVP5/pxBMryXvZbYg6oAuSTJRBIgAZ8IDOzZRS1n2OfXl08QeFIHApj5u/tKLn92AGORBDg4tHKIix3FVfLnj7YKDJ9dSaJaXr/j/CEyID/d1eGITqMCFNHDx8aTAAmQAAlYlQCiIyC4KUJdWFHW7amQlz7bJq3HXC2iippZjpW7LhwqhbmpVmy+322iAuQ3QhZAAiRAAiRAAu0JQPnZv3+/ZZWfpd+UyMuf71LLcu3brX/KTInXlJ+87JOOivV0O71SAbLTaLIvJEACJEACYScAh7+Y+cGrFeWTNQfkneX7DJvWLTNRKT/DJDvN3jZkVIAMLwEeIAESIAESIAHvCGC5C8qPFSO7I67X60v2yOebjGMG9uqaKncqm5/UJO+8jXtHyRq5qQBZYxzYChIgARIggQgnAP92UH6c411aoVsIbfG3+dtl7e4Kw+YM7pEhPzh3sMDwORqEClA0jDL7SAIkQAIkEFQCmPnBbi/s+rKa1De1yv989I3sPGQceHVsvxy5Wfn5iVVb3qNFqABFy0iznyRAAiRAAkEhgJkfKD/w92M1qahpkmdVXK9DlQ2GTZs2vLsgvEVn5VsqmoQKUDSNNvtKAiRAAiQQUAL6spcVlZ8DysHhcx9+IwhxYSQXTyyUc8f2MDps63QqQLYeXnaOBEiABEggWASsPPOzZd9R+d9Pt0pTy3GX3VcO5uW6Gf1l8uCuLo9HQyIVoGgY5QjuY+ux47Jme7F8U1QmDd+Gf5iowj8UdLGfV9IIHiY2nQSijoCu/CDGl9Xkyy0l8uriXaJCxbmUeBUkF6EthhXaK7SFy866SaQC5AYOD4WXwJrth+Sp15bJofL2hnvPvbVSZo/vJz++fJIkJ9p/q2Z4R4G1kwAJOBOA8lNcXCxWU34Qa+zNpbvlg5VFzk1u+5ymtrf/SG1zx3b3aBcqQNF+BVi0/1+s2yP3/eUz9QvG9U+Yz77eJXsOVcqf7jxXkhKoBFl0GNksErAdAava/GC2/H8+2CjLthwyZN41I1HuvGCo5KpXiopwTwgkYDUCldX18t9/m2+o/Ojt3XGgQl58f7X+ka8kQAIkEFQC+rKX1Qyeaxtb5E/vbXar/PTtniY/v3wElR+HK4QKkAMMvrUGgX/PWyN16gttRt5ftk1q6q3nd8NM25mHBEggcgjoHp6ttux1WG1v/+0bG9z6+BmjfPzcc9EwSaXJQLsLjgpQOxz8YAUCyzbsMd2MY8rKb/1OY7fupgtiRhIgARIwIKA7ObTazM/WA0flybc2SFm18Y/As0fly22zB0qcMnymtCdAG6D2PPjJAgTKjtZ51YqKGuvtwvCqA8xMAiRgWQJWVX60nV5f7JbjBlu94NPw6il9ZPqIPMuyDXfDqACFewRYfwcCGalJKq2yQ7pRQlqyvSMWG/Wb6SRAAsEloEd1t1JsL2wMefurIpm3rtiw8wlqtudWNeszone2YR4eEKECxKvAcgTGDu4pG3Yaf7mdGzyibzfnJH4mARIgAb8IWFH5aWw+Ji/N2y6biox/IGanJsjt5w2WHl1S/Op/NJzMRcFoGOUI6+M1s0dLDNyUmpCpp/WSLhnJJnIyCwmQAAmYI9DS0iL79++3VFT30qpGeeLNDW6Vn77d0+XBa8dS+TE3zNwGb5ITs4WQQM9uWXLnFZM91piTniR3XjbRYz5mIAESIAGzBLDcBeUHSpBVZNvBKm2nl7uApmP65sgD106QjBSaBJgdNy6BmSXFfCEl8J2zThM5cUz+8s7XKpbNsQ51D+qZIw/dNF1yOPvTgQ0TSIAEfCPQ1NSkRXXH8pdV5IvNh+U/S/YYGjujnXPGFAiCmsbHxahZq473S6v0xWrtoAJktRFhe9oIXHTGYJkyspcsWLVbxQIr/TYWWKpMGtZDJql4YJ2wzYFCAiRAAgEggC3uBw4cUIqG6+ChAajCqyKOKc/Or325R77YXGJ4XqwyFbhuRj+ZNCh6A5oawjFxgAqQCUjMEj4C2WlJcuWMYeFrAGsmARKwPYG6ujotthdiaVlBquub5YVPt7l1bpiuYnr94NzBAg/PFN8IUAHyjRvPIgESIAESsAGBmpoaOXz4sFhF+Sk6UivPf7JVKmubDen2VDu8sNMrS+34ovhOgAqQ7+x4JgmQAAmQQAQTqKqqkpIS4yWmUHdtxbZS+dfnO6X1mPFMFMJa3DSzv2bvE+r22a0+KkB2G1H2hwRIgARIwCOBiooKKSsr85gvFBkQ0ufNZXtl4QbjSO5oxwXje8r542j/GKgxoQIUKJIG5cTFxUlSEjwbB0ZiY08NWVpammUM9gLTu1OlJCcnS0KC/aZ3O3c+5XorKyvrVIdt8k7vX0ZGhk16dKobjkb38fHxYsfxw/0FY2jHvsXExLQNJmZ+6uvrBfeZcEtVXZP8+YONgrheRpKodnd9//xhMra/e2NnXKN45jj21ajMUKcnJib6dV05PvsyMzMD0vxTT9OAFMdCSIAESIAESMCaBGDns2/fPiktLbVEA3cWV8mz721Q9j7GwUy7ZSbJ3ZeOkoKcyPXsDIUlPz/fEswdG0EFyJFGEN7DmRaM7AIl6enpomvCKNdKMWoC1ce8vDzt1xl+odlNunbt2vbrrLLS2J19pPYbs3bZ2dmCX9hW2U4cKJb43uXm5mrF4Xt39KjxL/ZA1RnqcjBzhxkEO16bmO3Bkpd+3wy3o8PFmw7L62qbO5a/jGRYYabccvZASU7opN0TjfLp6VgVQL/gz8gKgpmobt26SWpqqlRXV/vVpJycHMHMKwTfPd1oHc8LX4UKkK/keB4JkAAJkEBEEEBE971797Y9NMPZ6ObWY/Lq4t2yXBk8u5NzRivnhpMKpXOE+juD0gPlx4rLcTp3KkA6Cb6SAAmQAAnYjgBmRODgELZNsEMJpxypapD//WSbHCw3nt1OiOssN84cINjtFYkCzlB8MBtldaECZPURYvtIgARIgAR8IgDvzsXFxYLQFuHeVLF2V7n8Y+FOaXQR2kfvHOx9fjBnkORlh984W2+TN6+Y9cEyv26m4c254chLBSgc1H2sc/3Ow7Jy2zo5VFErCXGx0qtrmkwaki+b95bKhl0lUl3fJPCcPHZQnkw9rbfExpzaceRjlR1Oq1F1LFyzRzbtLpHaxhbJVbG4xg8pkMnDe6oI7ubqW72tWF6Zv0EOHKlW69/HVRkpcuEZA+X80wd1qC9YCaj3yw37ZJVqS1lVvaQmxcuIvl1l5pi+2nt/6n350/Xy7tKt2nggWEdmapJcMX2oek2ULftXSUV1g5Y2sCBD1ddHEuPD+zUEi6Ub98vXWw+2sRjep6vMUm1LTbbfTjx/xtbMufXqe7FwzW7tO1nT0Kx9J8cNylffyV5qOcDcd8RMPczjngC8Ox86dCjstmiw8Xnrq72yYP0htw0e1Tdb+fcZoO4Hp3aruT3BQgexzAX7ONioRpJ0UoZExhZYkdQTi7YVxoT4FeKPVNU1yuMvL5GV3xw0XUx+lzR58MZpMrBnF9PneMqIm/qf5i6XWnVTd5Y+eZlacNLCbsbbE1tbj8tdz36sxfVyPh+fEd39f+69QEYO6a8Z0QbLCHrPoUr57//7XPaVVHVoRlpyvPzk6jO0h1WHgx4SSpRi+t3fviMNzeYDKUJhve+6MwUPyHDI3sNH5WHFAq/OAqXwnqtOlxmj+zgfMvysG0HDuZydjaAbGhpcGkEv27RffvfvLwVbm52lR2669p3s38O6Sxu6EbRV/OM4MzT7GUb4R44caWfzg2tTXwLDvSUURtDY3fXSZ9tl12HjjTAw8blEBTI9Z0wPs90zzIdlJxjoh9IIOlSzPo5G0I6eu/0xgubPEcNLyRoHGppa5CfPfeqV8oOWF5fVyN3PfiK7iisC0pH5q3bJo//8wqXygwr2HDoqdz79kVavUYW3PvmuofKDc8rVzMiNj70t1bX+KYxG9SMdSs+PVTtdKT84XqNi8Pz674tk8bq9+GhaGpRSeN2jb3ql/KDwipoGue9/5wlmxUItB0urNRaulB+0BYruI/9YLAtW7w510yKyvqUb98kDLy1wqfygQwcU77ue+Vh9V+y3+89KA4Yt7lDAw/3bfvO+Snn09fVulZ80Fc/rrguHBkT5CfUYwNYHyge2t0fKkpczIypAzkQs9vkfn6yT3T7eMBvVTMQTr3zp943gqFJI/vj6Vx7JQHl46rWlLvO9/Nl62Xek44yLc2a0+c4/vOmcHLDPv/vPUqlTSxSe5PcqH5b7zMqdambL3XZWd+UcV1Pkv31liTR5MXPkrjyzx37/2jJDhdaxjD+osa8KolLqWFekvq9rbFYzP0vF03Q6ZgefeNX/72Skcgpmu6HwYMkr3Fv4cR94+6siefaDb9S9xng2uJ8KYnr/VafJ4B7Gs+bB5OVP2Zhp6tOnT0QYOrvrJxUgd3TCfAzbJd9fts2vVuw8WKHZIvhTyMfLd5ie2Vi747DLX7hvLt5iuglfb96njBaPm85vNuP2/WWyec8RU9mhJH26cqepvMi0u9i/X/WY/fpiQ5Hp+vzNiJlB2JSZEcxCfrxih5msUZsHs2SwwTMj2/eXm74OzZTHPCLY5o6dXoH0ueYLVyx5/eGdTfLpWvfmCrNOy5N7LxkumSkn/dr4Ulc4zsFMT0FBgTbzY+Xt7WbZUAEySyoM+XYdqJCGJuNfEGabtH6XuQedUXkbvDwfBtnO4somwjmP/hm/opdtCrwysN5Fu/Q6Xb1uVIbeZsQb2yx35bni5i6/P8e8rcssC3/aFMnnesvT22sxktkEu+2weYF3Z9hlhVM27q2QR19zv+SlhbQ4Z5BceUYftWlEGf9EkMA+rFevXpKSErkeqZ1xh3f7iXNr+LkdgSqTvyjbneTiA5am/JFqL8+vdjIArVfLA95KSSWMBrt5e5rb/N4saaEgs/3ed6SjAbHbhhgcdOZmkC0gyTVOY+SpULMsPJVj1+NmZ3/0/nt7Lern8bU9ARgzY5t7OA3uW48d15a8FngIZNqzS4rcds5A6ZoRuNiQ7WkE5xO8L8OvTyBjWganpd6XSgXIe2YhOwO7ogIh2Gnkj2R72Y4cpy94cqL307yFXQO/Lu4tB7P5hxaeDI/gD2Oc6y1nf+rzti6zLPxpUySf6y0fb/lHMptgtR07vWDsHE4pOdqg7fLaX1bnthnThneXK87oLXER5gYBYW2w+8oxELDbjkbYQS6BWXjA+uZnSUaK/35Yxvq5xdrbLdpjBuR1oJqbad6xF6aGJw4r7FCGvwljBnq31Rz+lMzIUOUzJxCT2WMHmqvPTJs85QkWC0/12vW4t98Rf7+TduVotl/6Ti+z+YOR76utR+QxtcvLnfKDJa/bZg+Ua6b2jSjlB+4CsNzVpUsX2yo/uCaoAAXjmxGgMuFY8OqZw/0qbYx6qA7s6Z/fkdnj+0lWmjkX8vAZ0z2nowv0Wy8YY7of50wabDqvNxkLu2XIGSPMKVZQ2M4a29d08d4+AJ0L7qXadrpyJhkqgZ+oaaN6m6oOM5Gzx/c3lTdaM4FlXnaqqe6PH1wg/fKzTeVlpvYEsNR18ODBsO70amw+Jn+bt13z6tzkZrNGYW6KtstrbP/A+WJrTyPwn7C1HZ6cCwsLw+45O/C961giFaCOTCyVcuWMYYIbpi+CB9cvrp3iy6ntzklKiJMHbpjm0bN0gXqo3nXFpHbn6h/OHtffVD/Q5t/deZF+WsBf71WO/bp7eFDFx8bIAzdOl3jlbdusPPHD2ZKcGGc2e7t8OA/1mfWk3e5kPz7crcYqz4Wy6lhkXGxn+ZUa+3B7q3ZskxXfx+GauWm68tDu3osvFOufX3OGFbtg+TbBcSGMneHhOVyyWzk0fOS1dbJyR5nbJswYkSc/v2yE8nJv7oej28JCdBBb23v37i2ZmYE3PwhRF7yuJubXSrw+iyeYJgAv0IhD46sgEvB09esSu8G2qW3czn67sVzknIa6sJzy2PfOktzMwFjsY1YHS0Kb1DZyVzu6MLPy6K2zJEOFezCSs8f1k5LKWsHWfFcypFeuvPDzCyVLfQHhyTQYnlqhzM1QIR6KlENEOAJ0lt7dM7V+DO3tvV3P1UpZna+2Q7vylO1cj/55kJqde1yNU5+8LD0pZK9QamapWa79yj/TfhWWxFkwK/XILbNkeF/zxujYJgtjSTykwu2IzrkmertxAABAAElEQVQ//n7Gr2N9Bwy+084e3ruosDCThvaQLUVlUqkcXDrLuMH58thtZwnyWVWw9IHtzcHywu5rv7HDC9vc/bmX4trEHwT3Fm8Mp+Gr68NVB+T/Fu6Q+qZjht1ISYyV284eKLNOy1fBVwOxMG5YlcsD8HYNlwD4MytxcXHatnbY++Aat6okJye3RZavra1ta6Y/QVcZCqMNY3DeBCIUht6y0qN1snZXmVIi6tXsRIz07JIso/t3E/gVwTZc7CyBcSXsO/xd9tLrdH5F3Kh1ytcP/OngQd9F/aLFDJU3D/CK6np5dd5GTaFrVTeWnipEwFVKedBDBMC7KAwcg30Thi+cVVuLpRyxwFQIjOF9usmo/t39vnFt3FMiL76/WopLa9T6uQiW3m6/ZLy0Hjsh24urtVhgGSo+WL/uKapOZT+ETGEWeCfWYoEdPRkXbbiKi3aaYuHtrFS0h8LAMELxw/fR8TuJJVL9+g7zULut3oqhMAJl7OxrKIyy6kb5q1ry2lNy6qHrCuLAgnT57lkDw+rbx9tQGFlZWZqRs5UVH511MEJhUAHS6QbpNZAKEJqIYHP6r9Dy8nIt7kuQmh62YkOlAIWjg1hfxy9sPCQRz8ZuQgUoskfUSgoQviOI5wUFKBDiiwIEQ+f/LNktTS3Gjlkx0XPB+J4yZ2wPwYx9OMWsAoRZWmxtxxb3SJFgKEDmjRwihRLbSQIkQAIkENEEsNSFsBbhcm5Y29AiLy/eJet2u16u1+HmpCWoWZ8B0i8vXU+y9CuWABG13Z9lI0t30MvGUQHyEhizkwAJkAAJBI8AbKvg3NAfex9/WgePzv9atEuqlRLkTiYM7KJtb09StnSRIDBuxiyKHUJYBIp3ZIxcoHrLckiABEiABCxLoLq6OmyR3JtajsncpXvlyy3unSsmxsfItdP6yoQB3m+UCAd4GLZjuQtLgJT2BKgAtefBTyRAAiRAAiEmEGh7H2+bv/NQtfzfgh1SVu0+oG3/vDTN0DlbLX1ZXTDTg+Uu2I1SXBOgAuSaC1NJgARIgARCQABLXVjycnYrEIKqpUU5Mnx3xT5ZsL5YEITZSLCl/UJl6HzOmIKwGzobtVFPx65SLnfpNNy/UgFyz4dHSYAESIAEgkQAri5g7OyN35pANWVPSY2a9dkpiOflTvKykuS7yrcPgplaXWDcDOUnXPZTVufj3D4qQM5E+JkESIAESCDoBCoqKgSuPLD8FUpB9PY3l+6Wj1fvV3W7r3nWaXlyyaRelo/jpe/u6tevn+aE1NFRoPseRvdRKkDRPf7sPQmQAAmElABme+ADKxwhLXYVH5XnP9woB8rcOzXMSo2Xm2YNkEEFGSFl421lWO6CB2f8WcGhqrftD3d+KkDhHgHWTwIkQAJRQgB2PljyCkaYG3cINVuflbvkk9X7PM76TB7cVa6c0lusvr0dy12I1o5QFhTfCFAB8o2bZc46WFYt7365VTYq1/vVCIWRlqRiduXLxVMGqwjuSabauX7nYfl4xQ7ZcUB5llZbQfNUUNMzhhfKuZMGCAKDOsqXG4rk7x+vleKyGhXa4bgWJBMxvO5RQUadA2uWqRAT7yz5Rtaq0BlHaxskIyVRC69wXM07f7Fur1SoeEmYgk5NipdJw3rI3ZdPUp5J21+S21X8sw+WbZdv9pVJQ2OLdM1KkQlDCrR6l285oMXzQiDKfgXZMkHFWtpVXCnoD1ig/2NVWJDu2Smy4puDsvfwUa0+hKaYqMqoqG6QVdtUKAwVmiMtKUHFvOoqgwu7yOrth2SHCi/S2NKqRfhGnLPzwMIpOGqRKu+9Zdtk8+6TYUFyMpK0sCAXnTFY0lNCs0vkmBqDT7/eJUvW75WDakxiYzqrSONZMnuCueCzjmMb6e/BArHYFqtr64CK8wYWfRULxKCbqGJ0OctXm/fLvFW7ZI+6ZnBN9lAhWRDVHfHRvA0B4lw2P3ckAI/O8Owc6iWvXWqH1z8XwdansWOjHFIykuPkuhn9ZUSv0Mflc2iGx7fYzg6P8vDmTPGPAENh+MfP49nBDIXxl7mfy5/fWi7HVDwtZ8Gvl1/8vzNl6mm9nA+1fW5uPSZ/eG2ZfKYeoK4kXylCj9wysy3O1z3PfaIpF67yIu3WC8bKtWeN0A7jwfKUKhsKlVmBsvXED86WOWeOlorKSvnjv7+QNxZvMXt6UPMhgjyCveKBCnn5s/VKEVzn8mYOhe5XN0xVilrHh24gQ2FACf3VSws0xc5V588cWSj3qWsAAWBDJeEKhXG4vEYe+OtCTQF21dfJw3vKL6+bKsmJcVoMu0f/uVhWKqXYlQzskSMPq+seyraj6HYWSIOH4qNHjzoetsX7YITCQNDRkhIVq7CmJqSM4NcHO7wWbTjkdocXGjVhQBe5+sy+gmCmVhVcf3BkiDEyEvj7wdKiHW2AghEKg9Hgja6kAKVjyjeQFvl4wCB+y2vz1soTL3+uzWi4aipmZxarWQFENS/o4toPxOMvL9F+Mbs6H2k19c3yxfoimamipz/0t0VaEFSjvEhfo2ZOclWk68Mq4vsj/1jsUjFzdz4UOShOF0wZJi+8s0xeX7TJXfaQHkPgV8wsIJL8B2rW56UP1xjWD8Xyc5V39IDu6iGa2i4f4rjpgQf9uUlV1TXKj5/+SJv1aVeBw4d9KuL9roMV2oxGqOwDcJMOdTR4BAH+8TMfyz4V1d5IEO0eQYOnj+4tv3xxgXatGuUtVzODmF08e3w/Leiwns9TNHg9XyS/BjoafFNTkxbFPdQhLbbsOyrPffiNbNnvXklNS4qTm1Uoi/PG9VSz3daMhK7b+SBGoqdZn9TUVG15sbm5OZIvQ5dtD0Y0eCpALlEHLjEYClB1fYv84InXtSUoTy3F0hiWw5yn9FeoG/xLHxg/xPVyG5tbZeeBCm0ZS09z9/r1toMqwvpBFTzQ/MyPY3lYEvti7U5ZurHIMdkS79Gn/Uqp+EgtF2LJxJ3g+Ja9pYLlMEflI1AK0PPvfu32Ia63DctiPbpmtM1c6enBeg2HAvTC+6u1SPae+lSsZolKVbT7pRv3ecqqLaFiDBHFXRcqQDoJc6+YIQv1FvdatUz+6uLd8uZXRdLQ7P4ehFmfO84fIr1y2/9IMde70OSCE8OCggKBYuN4HzGqnQqQERnX6dZUeV23lanfEvhw6WZpaHIfp0aHVVJZp9m56J/11w++2q6/9fi6TtnUmBUYG1bV+ffrY1+J+19tZtsSjHwrlXKH2TUzUqSUpU17jpjJ6lWeZmWb9OnKnabPwYyVXaVFzbZ9usI8C8zimZWP1HfkmFq+oXhHALu8oPiE2t7n6x2l8utX18rybaVuG5yVmiB3XjhM8+2TqpZErSiY6SksLJTu3bsLflRQgkOAZIPDNailbthZ7FX5W4vK5PRhPduds7XI/U2iXWZ+8JnA1n2lMrJfN5/Pd3Xi3sNVXs2wYfztKljaqjf5YwAMMKNpVmrUsifsrHqqGTSKOQJY6oLyE8hlf081l1U3yr+/2C2b1bKXJ5k2skCunzVEOh1vCflONE9tw3Hs6EL4CszkUIJPgApQ8BkHvIZ6Nc3rjbiaLWpoMv8g8KYu5m1PIBicXY1n+1rbf4JNEmYynJdB2+eKzE/esvC2l8EYP2/bECn54diwrCx0yjZsBhHC4v2v92shLdxxylazPjedPUjbpYl89cqMwEqCuF26gbOZpS4rtT2S20IFKAJHL9/AqNmoK93UDiZnwQ4XbAunBJeA806iQNTWzWl3kqcyuyjDdDsqP+h3MPg68gx2+Y51Rep7zPZg1ieUhs4IY/Hy57vkYHm9W2yd1NEZI/Pk4omFkp6a7DZvOA7qBs5ZWVltmyPC0Y5orZM2QBE48tPH9veq1a58oDgvibkrMDGuvS8gd3kDcSxGBR60qqR4YTOAAIoTBhcEvCvdc9Kkd/dM0+XCx5JdJTczRfMBZbZ/+TkdfwwYnTu4VxfJTE00Osx0RQA7GYuKlMGxWvoKhdSrmWssdz355kaPyk9+drL8/PIRctWUPpIQ4nuYGRbYzt6nTx9t5kffGWrmPOYJHAEqQIFjGbKSzhzVT0YNMPdghRM4OHhzliumD9V8ojinu/r8/YvGa07lXB1zTuuvHBLCcaA/ct254wQPNivKLeePUX6RzCkfF04eJDlq9iUYctO5o0wVC99K18w66ZvJ1AkRmOmmOeZYxKltzj++YpLpa9lsuRGIzO8m6759EMU9VIFMVyjj5oeUkfPiTYfd+vWJReT2CT3l/itHSp9uaX73NdAFwINz7969BT57aOAcaLrelUcFyDtelsn9h7sv8Tj9P6BHttylbviuBF6SH7pxuuCh4E7gAfniMwfLf988QzzNy2B25A93zpGffecM6aW8LfsiQ9Sv7vtuOEue+OEc0wqaL/X4cs5s5RfmkjOHyK8VC08zA8P7dJUfXjzOl2pMnTP1tN5y1YxhbvNiJu2/rjuzg4dutydF4EEo3LoDTqPmd1Yxk35x7RTNOeVPrp6sthQb5TyZfqNSqlw5snR/VnQchWuPffv2CTw7h0IOVdTLH97ZJH9fsENqGtzb7gzMT5cHvjNKzld+fWKUJ3ArCfzYYGcX/PnAlxsl/AToByjIYxAMP0D48qQlJ8iUYfkCD7jOtjwIAQDfP7+8/qTnW6MuFqiZIYSEQPgIhK1wlLTkePn+ReMEMx4Q7IQZp0JNLFfhA1ztpEE4jBd+dqHypBovCcoLNZzI1dQ1yc7iig7OGvEwcvaigwcSlK3f3HaW4BdSelKsTFT1wbndofL2gQuhtGFmw3k7Otrs7Hkaq2nw/ovt+Y6SrLwjw8+LszsfV2XAs/NtF46V25Sna6zZI6QHnEMeVu1ydr6Htl05fZjywj2lQ+gM1B8oP0Aoa5xaXoONCnZ5NTjtbkJokAdvmh7yh3g4/ACBxRgV8iRPLW99o3Y3Ohsu983LkgdunNa2ExKzlCP6dtMcIx6tbR8eAfZSCOty6dQhKLadRLsfIISwgKEzApmGYtYHfrfeW7lf/m/hTimrbmo3Fs4fUhJi5ZppfbXlLqOt7bg29RkXxCLDLFYoBM4lsZ0dRs56/cGql36AvCPLUBje8fI6dzBDYZSXq9hdyuNnhYpltVHFo4JHXMzsnKa2XacqBckbQVwrLRaY2jHUPTtNPSC6qtkh17Y/cPC3cM1uqVYKDpSoCycPlOx010s9yLNhF2KBNar4WInalnDMnixau0fWKs/RLcqnDkIPnD8ZccdO2uTjFxJ+XdbXn1TKsBUZ28mx+62b8qyMbeVQ8uBjB7HQtFhg+dmasz/EF9u0W8VFU/WCBfJCgYEX4JOxwE4IYoENLszVFLn1qm3lSvkDr2HKazaW3uCHaLty/tiklAo8VPGwNGJRerRONisetYo9Hp4j+3V3O3MVyFAY+vgiBtZJFogF1kmziemneIRDwhUKQ+8rdrtt3lP6bSywTgLlp7+6vowE1/zub2OB9VTX8hB1DRgZjOPhhS3KENi8RFMoDCgMMHTGD7pQyOqdZTJ36V45asKn2KRBuXLF5N7qe+7epw+uTSgjENxbgh2QFT9UEaw0lFvaGQrDu6uTCpB3vLzOHQoFyOtGWfwEZwXI4s31qnnBUIC8akCQM4dbAQpm96JVAYKih+3toZgxKVbLXa8t2SPbDnpeXuuelSTXTO0rgwrMLbeHSgGCLx8oPpjJDrVQAfKOOLfBe8eLuUmABEggKghgezuWu/SZ2GB2ulGFrXj/65OBS13Edm5XNZaZzxvbQ2aPyreUnQ8U5OzsbC1YKX35tBsyy36gAmTZoWHDSIAESCA8BKqrq7VQFsGe9YFd0Vdqd9c7KnZXtQcDZ5AY2TtLRW1XW8fTrOOegE4Mw3ONBqJWKkCBoMgySIAESMAGBDDrc/DgQW3mJ9jd2X24Rl77co8UHWm/ycFVvTlpCXK18uczsk94bNtctQlG8TBszszMNBWo1FUZTAsvASpA4eXP2kmABEjAEgRqamoEfn2CLTBsfuurvbJyu+ewGTDqP2d0gcwZ08Ojy45gt1svH4oPlrqg+NCBoU4lMl+pAEXmuLHVJEACJBAQApj1QeR2eHVGFHIs6QRDEJNu3rpi+XTNQWl2ckvhqr7T1HLXlWrWp0u6NZa7wAUhK6j4uBqtyEyjAhSZ48ZWkwAJkIDfBOBuorS0NKg7vGDng9met5cXmdrW3jUjUfPnM7xXlt/9C0QBVHwCQdGaZVABsua4sFUkQAIkEDQCmPUpKSmRurq6oNWBgnceqpY3lD+fvSbsfBLiOmsenGeq4KXw8xVuoeIT7hEIfv1RrwBhi+eLL77olvT5558v/fv3d5uHB0mABEggEgiEwq9PaVWjNuOzZle5KSSnD+4ql05SEduVN/dwCxQffTs7bXzCPRrBrT/qFaCdO3fKG2+84ZbyqFGjLKUAIRRFvIsYXggN4fzL6bhyqoF/zt5tXec9Lq1qbT5ehbJwFHgadhVXB+1IdMrreF6g37uqr1m1IVaxcL5RucprxMJVXrTdqN9m+wWvxJ3UP0SF9yRYJkD7XHH2dK5+HO1FXc4+SFyNtX6O86s3eZ3P1T+3KFsPivUIwGs8Zn2CGbm9Vnlr//Dr/bJ4c4l2PXui0KdbqnznzL7Sq2uqp6xBP6778UlPT+9wPwl65awgLATaP+nC0oTwVrpjxw6tAWPHjpWZM2e6bMyAAQNcpocycdfBcnny1aWy+1ClHPvWUxiUj4E9c7TdEXD/jwc5YlmdpsIxZKcnyfqdh2X/kWqtmT26pmshMqpUiIh1Ow5LtQrdgPMR/gEPvW0qVATOhyBWV0+Vv1e3TNmiYishTlicmpJGXSkqrMQ3KvRDTUOzlheP9u7ZqfLDS8bLlJG9tLRA/QcfJM+8tVIWrNolderGCkHMMITDQKDPwxW1bSwS4mIkv0uaCgvSIOijLghPgQCr4KDF7VJByApy07QQCd/sK5PSyrq2uGT49Tm0d1ftfIRIAJfczGQ5Y3ihXHPWCFMR6msVl7mLNsvn6/bIwdIa1WCRQhVH7axxfeWyqUP1ZmmvUBTeX7pNPvt6l4rHVqHFJgPLMxXH76gI7p4CrqIQhBh5beEm+WL9Xo0Hxq5vfpYa164qhlqNFiKlvqlFxWiL0+JlXT1zmNZHx4bsUdfUv+dvlK+3HtT6DpYI5orArwg0akYQI+6tJd+oMCRHtOsoPSVBxg/Kl6tVPxB7ixI+AnoML8TxwvtgyNGaRvn34h2yaX+NHDNRRVZqvFx6ei8Z379LB4U9GO1zVyY8N2PGB4qP848Hd+fxWOQTiPpQGL/97W/lww8/lJ/97Gdy0UUXBXxEAxEKY+6iTfKXd1cFvG2BLnCWChB6/w3T/C4WoTD27D8k1zz073bKjN8F+1FAklIWEVx04tAehqXsPFghv3xhfofAsvoJPVSsqRf+62rpU9BFjlTUyK2/eVXFX6vQD7d7hSL76K2ztDhk7Q44fEDMs1+9tFBTZh2SPb694ZzT5KZzR2v53v1yqzz71grDX+sY01/8vzM7zCzqlWDW6fevLZNPV+7Uk9q9QmG9/ZIJcvm09spfu0wR8iESQ2FgtgezPpj9MSP6LjDsCDMjCCj80ddFamdXsbSYiC2aoGZrz1Fb2s9WXpzh0TmU4hwKA8oO/PiEI2RFsPrNUBjekQ3tFehd20KSG0tgkEGDBoWkPm8rWb5lf0QoP+jXgjV75KUPVnvbRZf5b3x0rmWUHzQQ0dYf+tsi2almhlwJgrDe9/w8Q+UH5xworZbvPzFXjtbUyw/Uq5Hyg7w19c3yX0qZOqjOcSUIEIvjmMnzVv756Xp5R83WLNlQJE+/sdxQ+UG5GNPnlIJkJM+/t8pQ+cE5mHD489srteC5RmUwPfAEEK0dYSz2799vWvnxthUb9lbIf7+6Vj5YZU75mTw4Vx6+boycNy68Pn1SUlKkZ8+e0rt3b1spP96OH/Or1Y5ohoCdELt371Y2JLHSt29fDQVcwDc1ef9QCRZHLHtFkvxnwaa2pTRf2/3CO18JoqxbTeDH5Dn1MHcl//h4nUAJ8iT7VaT5nz/3vlpGLPGUVerVst8L77ue+XtRKZr6sqDHglxkwPlPz13u4kjHpPfUMp0rxW/v4aPy1uItHU9wkQIlqOnbJVYXh5kUQAK4h+3du1fwGgyBB+ffv71R/uejrVKijJ09SbIK0t5LxSsd1StNMsJo5IyZHmxmgUmDnWZ9PPHncWMCUW0DVFRUJC0tLVJYWCivvvqqvP3221JeXq4ZwCHt+uuvl9mzZxvT+/bIokWLlMGsa8NP/MrA+rIvslfZZsDGI5IEU+LvL9sh150zyudmvz5/rc/nBvvEDbtKpFJ5su2efSrSc6sa+4VrdpuuevmmItN5l25Sv+DVpZWWnNB2Tp2yM/pSzd74Iw1NrYI/s7JwbZEM65ffLvuitXvb7KfaHXDxoVLZiKzfXSpnntbbxdHISMLuIF3wHksqVhL8cNODl2J5Bz/svBXdBsbVuQfK6uTtr/bI+j2ul22d64pXuHKTRVK+3dj1TVGZjBlkvITsfH4gPqM/UHYQnT0jI0OSk1WDlKB/Vhu/QPTXzn1z3OiCsQuEPZv335BAjZIFytENoPft2ycvvfSS9iWBMTR+PeHvkUcekTVr1sh9993ntrV33XWX4awRbIwuvfRSt+cbHXz/K9d2FUb5rZK+VRlU+6r0oQ+lR83ZH4SrvyVVzTK0/ymldt/hSq9mY7zZJYWdYRX1x6RXj1P1HdhZ3Gb8HSoGRcqI3HlM9x1RRt5eyMGK+g5leHG6pbLGx8dbpi/YLAA7H3hzxgMfSzz+imMZJWqjwNwlO2XZ5mJTCi9Me3KSRNKVfqia0yalajOFY7ltB4LwBg9LXK9du3ZVu1o7bq1PTQ3/rrMgdFsrEnZc+LOzwCN3ICSqFSDd/ge/EB577DHBdncINEvMBv3pT3/SDKQnTZok06dP146F8j99V1Yo6wxEXVgq8kdgN2Jlwe4wR3H+7HgsEO/hmsBRvFGgHM/z572rOr3tt3M//GkPzz1JAJ6cDxw4oM1kB5pJhZq1e+vLnbJo/QFth6Kn8uHtIUtFrchSz15Xnh+OO31vPJXny3HM7OTm5mrGza5msXwpk+fYl0BUK0A33nijzJo1S5sazc8/Nb2PX1GXXXaZNgsERejll192qwB997vfVf5zXC8nYCnN7I4K58tsSOGpX/3Ox6z8ubfaPu9rn9EvbKEuV78WrSq56qetY//SEztru6TMKgTYqo6lQrPSJS2+XX056nOoJT8nrV0bUH9+jne/ortlJXcoI9T98Kc+zCroSyj4vjc2hm95Gru6Dh06JAhgGkjBlvBqZYD/9lKl+Gwo1lxBeCofkzyZSvHJVoqPOwfOWco1R7CYYUlEX+bCOLmqB33Tl71w3Oie7am/Vj6OGTZcGzDtsJvoOxTRL8f7rz+zeVGtAGHmZ8iQIYbXyVlnnaXNBO3Zs0ftlIGTOdc243fffbdhGdgG7+tNakB+hsQrnyzNLf7NqBg2LkgHLj5jgM99RpNmjh0gcxeuD1Lr/CsWW9m7ZSR06N+EIQWyTNnrmJGBhbmyteiImazKZ0+uJMaeaFdfgroM4adn0x5zZbiqCEsTcbHmr63Jw/LbtQFlThqar3webXRVfIe0eFXXaX1zOpTRIaOFEzCjoCtAeMD4+r32p4u4D8GfD+4rgbCBcGxLbUOLLNhYIgvWHTAVrBSKT25arKTEtCo/YY4luX7fPz/b0FTA9RmeU/FQxFKXvrTmLrQHHpSOClAwHUJ6bnlwcuD6hALkqCAEp6bQl4qlTN0OD/3Tr39/FCDXT/TQ982SNcKnAgQXlLsvVjAbf4MfxsTBbJdR2ROVIgDng/7Ir26ZrSl+/pQRrHO/f9E4l0V/97wxhr5yHE9IToyXx++4QDmq9MwID5jbLhjreHrb+9suHNvOvqLtgMk3l0wZIvAHZEZG9e8uE4Z0NF4dPSBPpReYKUJzJJmRoqYJKD4T0Hd3BdqhIXYTvqMCld7/r9Xy8ap9ppSf0/pkywPfGSW3nD3QlPKTkhQnowfm+dx35xPx4xWz69jOris/znn4mQQ8EYhqBWju3Lny/PPPa0tdrkDBsBACD6H4woVDrlUeiKeY9MYbjvY51lmgPDH/5rZZjkk+vY9Xv7SfuftCzduzTwUE6aRbzh9j6BkZ3pf/6zpjh4FoEjxv//Hui2Vwr27y559dIanKq7aRQPm568pJcppSPlzJiL7d5J4rT/dJCYLS8oOLx8k1ykvz7PH9XBXfltYnL1NzANmW4PTml9dN9ejpeaZypnj9bHPKllPx/KgIYLkG/nywwyuQyzZQfN5dUSS/VIrPJ2sOSpOTrZkr+IN7ZMh9l4+QH547WPKzk6VX90yZPrq3q6xtaUkJsQKF29+wOfj1D+NXuCyBs9TERCrUbZD5xicCMb9W4tOZNjjpmWeekYULF2pT2dOnT+/QozfffFM2bdokY8aMkbPPPrvDcTMJgVhrnqEeIMnqJrJZLXk425ngIYpYWC0ONy+EwUC4iGoVEkK3NMGSB25WCGmB0Ai6IERCQlys+tXXfpkNXloz1S92OADUJS6mk2af04h92Q6CsBTnThwgv7/9HG0XisMhn95C2cxSe2enjugp63Ye6uAKAPUhtpmzHQ3io+lhQvSK0W/wcDQo19Oa1NKisykOwk+0KhbfRhvRikEIkJ9fM0XmqD66kz55WZqnaISgQJgOXRCf6/RhPeWhm6bLpJH9taXUbmob/fiBuVKpfAchTIdjX7C89cvrp8pUD1vGB/bsosJb5GnOEo+onTq6IB5cbzXWrWq5xHH5FDNzN507Su64bIKarYrRxgrhS7plpcge5dMHzhd1wXV12bQh8l/KC7TjFnz9uP6aoJS6s5USBaZ7iivbXUcI64EZMyiOZuKh6WVa9RVL4PpsQyhsgOBao7S0VNvdFUjFB/G6Plp1QP46b7tsPVDd4Tvjij9idt08a4BcML6nCtHSfvt/nrIPK1BLw3B1gFAwusSo+8Wgwi5ywemDJMePWWHY7sCwGTPyWO4wMkXQ6zV6xRKK4xJYIJka1RnqdPDB8ixWLewmWN5zXALT++fP5ERUh8L4z3/+I3/+85+1L9Tf/vY36dfv1K/h1atXy09/+lPtFxe2yPvqKToQoTD0gcbroaNNKm5XhYrJFSdDemRKamKs5sm3SDnYg8KDh31PFXsKUqO8BB9QD1cEQ+2Rm6EpL0iHd+FyFTML4RYKu2VoygT8DX299YA0NLaqGYpcLe4X8sIhIR7m+PXWS+WNh7KklKKvlIfqiupGwQzBKLUUEkjBrzvsbqmvP2kIXdvYLF9vOSBVqj8DeuRoca5QH+Kjbd5bJonKAGHc4HxtWQk7jeA9GwoB4pmN/3bppladu1/1G+vGuFnryzFrthdL0eEq7QY9Sc2MIBAsHPYVlVQp5aFVuqmHeG6m99uKq+oaNc4wqO+p6kv91o8PtuXiS4x24Bc9pEEppPtUfS1qlwzimWWnKWtSLwVOGOEdGgouxj9ZKbbYQo/4Z1VqbLNUmeCB9hjJwTJ1XVQ1aDNTGGtvA7Mi+OshtdW9U2yidDrW6LWRtFG7rJKu7zBCe2A/gqjqwRBcG7hvYKkLNj+Bkhpl4zN/XbF8vvGQqdke1NuzS4pcNKGnjOhtbkMGfFThXgIlHNccbBh9FSibmZmZbUqnr+Xo50E50B+WGDs72gAxFIY+2uZeo1oBwi+se+65R9auPel4Dz6A8Ift8XBuCPnhD38o11xzjTmaLnIFWgHCcpz+KxROG+2o6TsrQC6wRmySKwUoYjvjouH4hQ2jVCwfB/Lh7aKqkCeFQgGCYXVZWVlAd/EcVY47P1t7UJZsKWk3U+wOIJa3LlSKz2hluB5KgYIOh4VQfFz57/GnLVSA/KEX/nMRt02/JvDjUTeCxvPCV4nqXWD4Jf7kk0/KK6+8onmCxqwP/iAwrvvxj38s8AFEIQESIIFgEsBSORwZ4jVQUq6WpD5Vtj3LvlFL545rum4qyFOKz+VT+suwghS3s4VuivDpEB5sUHzw5+sSl08V86SoJhDVChBGHoZ0t9xyi8AnUHFxsTatDSM7/FqgkAAJkEAwCcBeAzM+gdxSj2XIT9WMz8rtpe1s2dz1I095L0SQ0inK7i5ObUII1TbqQC9zuesjj5GAM4GoV4B0IJjexrZK/FFIgARIIJgEsDyIJWzYouhT+f7Wt7ekRtvNtc5krC7UB8Xn/HE9ZUz/HOW9uZP25287PJ2PmXcs5WOZCwbOFBIIFwEqQOEiz3pJgASijkAwDJy3HjiqKT5bD1SZ5lmQkyznje0hY/rlhGypC7PtUHowu85lLtNDxYxBJEAFKIhwWTQJkAAJgAAUH93AORDbr7HDb82ucm2pa7+K0m5WeuWmyLlqxue03lkhUXyg6EDhgf8efQu62bYyHwkEmwAVoGATZvkkQAJRTQD2NLDzCcSOTfjrglHzPLWdvbymyTTXft3TNBufYYWBiaLtqWIYNWO2B0tdnO3xRIvHw0WAClC4yLNeEiABWxOAHysoPoHY2YUApYs3HZbP1V+d8tVlVob0zJBzx/SQgQUnfYOZPc+XfNjCDj872MmFGF0UErA6ASpAVh8hto8ESCCiCEDhgeKjO/L0p/HY0TV/fbGs2FZqeis7XF3Cf8+csQVSmBv83azcwu7PCPPccBKgAhRO+j7UXa+8Ih+qbBBE107o7N5L7JGjtfLJ8p1aLbPH9ZPuysswZP6qnbJq2yHpl58pV84YoaVVqGn6f360USpq6uWCyYPagl/OU3k/WrFDumakyB0XTVRT2ifd4L+/bJvm6XjsoDwZP/hkoMw9xRXynkpH2y6dMljVl66VvVaFs9ipvFfDa/S4wSeDZyI0BUJ7IFTD4F5dNK+xyIwwEpv2V0tCzAnBtD2mz+GwcsmGIilT3qtHq9hY/QpOOmeDB+US5fEZYTu6q9Afuufi7fvLNe/VhcrzcW8VngICb9LfqPqOKVuMocrTdfq3gTm3qLTNe0s1D8xnfBtzrbm1VbbsKdXc+sOVvydP0PA6jRAAORlJbR6m4U26pLJWs7OAN2kwMRKcu3TjPi1EyaShPQShBSDw5F1WVS9JCXFauAr8woYtCfqMviO0hbswFSgD58NDeGZaok8eplGGK6lX4RTQP4RR6ZadonkTR74jKq2stlVOtDRKuvI0HqkCZmAHj+sIK2NGmpqatJ1dgdhCDsNmeG3etM+8t2mEGxlSkCbnjClQMz7BXerSbXuwzBWKmFzwMl5SUSdN8M6uxgOeziNV9O9klvpOwls2JXwEotoTdCiwB8oT9LZ9ZfL3j9cqxaVYC3GAtiNe09lKsUFU7wwVw0qXNxdvlr+8u6otn55ulVfE7Gp2iF2GdiHUBm5yjjHNkI7QDggR4ShKD5BcpZCVV9e3xTFCTLOMlARN8XH0+aZOVw/puHbxz1CWqzYgHcqUcxsQMgRR2aEY6tKibDHeXLxF3lmyVY6ocCG69O6mXPerh+b2A+Vt5SSocADTRvWWm88bLSMG9W0LhfHZl+vkgb+qWHQO8ZNQDmKdZStlqrTyZCgQpCH+UqEqe78KbYGYS7oM650rN8wZpZTQU1HZYSALBfWNzzfLQRUeQxfEB7t65nA5Z0J/Pcnr1y17j8jfP1ora3ccbotfBj6Ig3ZI1VWqlAZd+hdky/87e6TWdz3N6q/LNu2Xlz9bL1vV900XKJoXK4X+O2eNlB75J73OOobCgG0PtrT768sH190K5btn4YZDUqxmfswKrnH8LslStwB1+WqC+IBjB+YbBu81KhtLV9imbqTEQdnBEheWukJh2wNFFOPx6cqdbd8TbNdHDLybzx0tQ9T1b1bC6Qka97b3l6rvpLpnIGSNLvhOfuf/t3cd8FVU2ftAIIWQkEYSIIFACL136Qgi6KKIrru69l3Xtmsvu3Zd26p/29p7W7uuYkFRBEF67wFCSQKEkEISUoAA+Z/v4jwmL/Pem/fevLySc/iFN+/OnXvPfPfNzJlzzz3fxD5MSuz5Nam1hU+hwtCj4XpbDCDXGHlVwwoDaBZ7YP7v40UODZq2ca3o8asnK7LTe16fQwv5Ji5iPQJgvb730vHK2/PPV35UniN3eoHB+uLt59HQXp3orW+W0OPvnaBbcacNR3VhBF/GDwQEyd7HRtXSzXscVVXs3Xcx2SoIZd2RmQuz6dnPlrAXyp2jiKax4XjT+ae4d5Afar/81XL6ZO5Ghz33YG/gW/ddxMZolOKRQuZm8HVVVFQ4PMbMjnLE96zfR/M37mNPpfn4npioFhTV/BjFhNfxWBr3BO8jjFCzYmQAaXl7ENDcmCu5djN3320v/qA8nkb640XoxvNOoWmjTr6YGNXTyvxlAIFT8G6+Jldk79VUafA5cXAXRTzsLWmwGEANoHVa0KTZ4J0iY9FOxAN4s+x1/Y5CuvfNn50+dDAdsYzJQuE5wBuGiG8Q2MWM6SB5/HLBZlq1tcDtTmCc/LxiG/XPak+3Pve128c7O2Dt9kI1jfcNe37mrt7lrCrhPEAaO7h7e6f19DtXsufxoXd+YVpd9wVTkvDwgeU+UAXG3RvfnuAEdKQjpi6ycwvZg5ZF+fn5tHv3bsK0l6eykxMXfrE4l96bt5227q1o4BV11C6SF844pRPVHCyjljwNzpe9QwE56YHKGkUi7LCSbgcSE8KzgwzVyNKclJSkvArYRrLYxhIYDTc8N4unxCuddrmU73t9u6TYpo2dVUaskmbAeXtfdtaP/b6nPllM89fm2hfX+76z4IDyqA70klgaRp6wwdeD1umXxvtFO1VDdjpC4GWeyjLzxl3AjO3/+WKpo2ak3CIE3vl+DR21m5Jzp2mwxF/z+GfuHGK67otfLKMKjhsyI5/y9Nj0MT1cxjdpbcE74onxox3/LuN25ogsat2K52oCTBCv9ca3q1xqBWLz7dt30tc/LiDEl3mSwRm/nZU5JfQzM7Ln7nf+cLdXqEdaGzptQHvCUvYlG/NV/Jx9HaPv2bnFNGFgZxVLZrRfXwYDAUYPMlU3psGj1wHbXy/aSvAAuRL8JvHbfPW2s1xV9ct+GDbw4JuRj+ZsoLNH9eBYwlZmqksdCxAQA8gCEH3VBIJrN+cWmW7ejKFkujGpaIiAN8aP1iDeyn0h5SaNH/SN8/iVA6/PGdPTpSqIO9q+94DLes4q1LCRgWk5uPoDTVayN+8gT0M5knA2fDhxMrXmmG44WzZsL1AGkKP6RuUHKg/zFFch/cqM7Adrao2qGJbBqzs0K4km9m9P6UnRtjoI3HdHVm8roJF9jGl+4PFBTA+muFJTUxU9BVax+VPmrd5puvucPaXKWEpre2LRhekDG6HivDW7TPeCa3Lhhjw6i40gkcZBQAygxsHZo17w4BERBHyFQF6hud9X/n7Xb+JmdMwL0N+zo+ssgg0fJke3GT7aORaXm/fcYDXXPI7vWbur1JQnV+ujdWQLGts7lcb1TaU2BqvpsBrUHcHLlF6wohDxPghoxvRWYwQ06/t3te3ubw5jGIgGUL7Ja0zDw+w1qdWXT+8QEAPIO/x8ejRuUiKCgK8QMBtw6SzGxB3dwgL094xVRXph20N5fKIdrLRupvxA+iPqb1cfPkqLs/eroObCspMr9urXMv7Wni2uU/u1o+Hd2qoVica1UFpfZ8f1TuzRThFTXPD0wOPjzykus/q6qqfttx9Drdzfn2avMU1Pd+trx8mnZwiIAeQZbo1yVEaq77O3NsqJSCcBiUBnXoJrRrBU1wrJ4DxQgSiaXq3Y4EnktCycxcCpJDvIC4SgZkxzrdhW3CB1g7MGYcr0Y26uCWz49EgzhxFSD5RyXiwzgowT7VNTqFOnTrYgYDPH+bMO8netzdlnWoVOFv1GTXdosqK7105Gqm/zN5lUu8lUEwMogIc6IbYVYVUA5u/NCOIFjumT4Jg5SOq4hQBy+hzm5I3eCHI2lVe65xkw0x/SIeChaOY3EM7noSV+dNV2KidmRL4hd+NO9O3iga3PVaTf589tBDJnpkRT73bhVMv5fMxIP17Fp8mhI8do2bYiWsCGjzukpDi+Fc+xjeqZQuP6pFJSLCfxcUP6Z6byaj/HcTLHODq4kk8HdGE1vLL++UlDKIJX4gWLTOJYMbMGEFYXpvJy/0AUBJ+/NWuNqYB5rJQc1Tc9EE8jZHVykD0iZM836E7smrOHuHCFnzilzPbxdN/l44Pu/IJN4b/NGMZJyzI9VjslIYY+eOBij493duANnBPl/Al9nFWx7bv09AFuZaG9dvowlQLA1oCbG1edNcTUKiQ3m/W4OlY5lZWV0a5du6i4aD9NHpxhqq0+me0431YCbS8op/fm5tAdby+nD37Z4Zbxk8YR1X8an0mPXTKEzh2Z4bbxA0UHciLA1naxQVgRBaNnL+fZ21FKVMihStUcb41Adzxcg0mmcLLObumJLlVGWoprzxnmsp6/KnTgwOzzxvUy1f1lUwfYMsmbOkAqeY2A5AHyGkLnDXibbwJeoMz2CYoqwdGbPdysj/71NOrJFA9l7FlA1mgR6xGYMbYnXTS5P9OEdOBVUSdWnrjTCzw0r/zjfMrqmEydeUxnL93izuEO6yJW7IbzRqis4AOyUpmOooq288oYR3IOn8efzxzkaLdhOehAOqa0ocWcZNPR79DwQC68lDNV/358b0e7G7UctCpITlpQUKCyHcMQgiDbM7InOwtC7ZDcRi0Rf/37jTRzyS5l9JjFAt7ZIZlJdOG4LjR9RCfqxBxdGnWLpwCAQmYjZ+auOHScSnk2bB8bPBXs8WGnlE1G9k6nf1w0xvbd1QYyPSPxoRU8Zq76crYfsTAjWPdVW/fWy3yuPwb0MnddMtZ0Pit/5QFC1up9nM9oh5OVlOeN76USmerPz5NtyQPkHmqSCdo9vNyubUUmaHSK9OlICY+ly+COguDtAm9K5/Ibhv4Nb3n2brr3jbleT9WoTiz+D/EO4K+q4lUs+odHYmyUoo7Q57EBDUYEr0GutFs2DLqKLhwjAJoHDYv2PE0Tz9MI2/ccIPCMaRLNEa2tIsOphJPYabOD0CE2Oly1q9cBgaLI1ozEkvpyuNdhYAxnni5NMHUye/l2lRRxCyf6g0DfPpkpFMNtrOPEhDBGIXi4ThrShS6Y2JcyM9JtVBgr12+h2176UdF3qIq//RfFb+uIaUAOEW26rQvHz3Ti+IBcTmK4g8shmI6DMQajLCut/tsyuNNAhbGROc2Os654oGDa5I9MhTGUj/FUsNrm/dnr1HJd4ATB6pustATawzxuW/NOYIE3c9z4L2T6iH6Mib8FdBW4FpG12Vn+HtB5LOXrBzgf43kkho4i2Cg4UteCdu6vdiu2B+ccz2vnx/RKpdG9ki3lRoOhgmDmcP587vPl9Asvt9Z+K6pf5pn6C9O3TB2e5Rb0WBWGZIj+XgavKY0cTf/jxKPIpaOtDAMP2Gjm7cPv3p2VX/7KBK2dy/y1u/ia3KSMVvyuYBT3Z27DP/J9YYgbSUm19ow+JRO0ESqOy8QAcoyNJXusMoA0ZXDTq61rrsg1a6oOEm7szgRviJDeGSez8JZXV9PyTXupD5dphKWos3FnIeUWltGkQZ0Jb0sQZLqduWgbgQagLz9ANQHRKrwMAzPbMxniSff60k35aqqjn65uDRtsOfsOUFd+gEexcaBJKXN54aatkX+iHBmKjzaPZE6jo9SCTnKAlbIBs4+TPXbvmKgMCK0NcAXBIAJhqCYwgEDU2S4RJKQndStizq5jnGsDMS2a1DB+OWw0deSHeBumONAExsvBmsPUgeu6WiIMXrAqNgZi2bDTr+Ko5Lw8zfgmF80GmCbJyck2A2jfvpNBnnuKypVBBvJVTfCgxvlFRbSgcCYd1QRZcmt4pREMSX1/2n79J3KLwEiEUeatx0HfrtKNzw9kqHrju3lYC2oZGU2HqsrdXKekb926bXgycA1WVdVfBu6qh9KKQ7SAc/Ys44DmEgTSuCEwsHt3jFPL2Pt0inc5RmabxgourN7Cg1y7PvXHgvAXhn6HxFjlzdLvM7sdaAaQXm9c1yBDxXXmyQpZfxtA2rnYrkmewnSXjkZrw9GnGECOkDEuFwPIGBfLSn1hACFvBwQEjK4MIMtOpBEbateuHZWXl/vdDe+LU3ZkAPmiL3+0iYd0QkICFRYWqmzC/tABxhnIPMHT5Q5VBchI1+4spUW8hH1zfpnbma9jePnYqJ7JNKZ3CiWyB8YKgaEDowd/RkaPFX3o2whkA0ivpyfbgWIAeaK72WPEADKL1Il6J18r3TtOagsCgoAgEFAIIL4Hgc34w7ZZwfJ1GD1Yvl6jD6Ax0QC8PT3T43iaK0UtZbfCy4YpKHh68cCGQSkiCAgCvkFADCDf4CqtCgKCQCMhAC8PPK0HDx50Gt+jV6eUp7WWbi2iJVuKqLDMXD4d/fHxrSNoRPe2KrbHCm+P5ukRo0ePsmwLAr5FQAwg3+IrrQsCgoAPENCmueDtqakxZ8Ac4niz1dtLlNGzZY85GhC96hzORYO6JtOE/unUK70NT695l8sJ3h1tWqYxprf05yLbgoAgQBxnKiIICAKCQJAgcPToURUfZnaaC6v5NnE8zzL29KxhPi7E+bgrKXFRKrZnJCct7JByYrWdp7F3WL2lGT2Y6hIRBAQB/yEgBpD/sJeeBQFBwCQCWM0FowfBzWZkx76DtIynuFbkFFPloZNpEcwcizqRnGJgcNdEGtkjmTLbnWAZd7Ua0KhtrFbSGz2BzL9lpL+UCQKhjIAYQKE8unJugkAQI4AkhVgNiD8zHpe9pZzegWkplnMwczEyAnog3Tu0UUbPwC4JnHqA6eA9EBg9rVq1Up4eeHuQWFBEEBAEAg8BMYACb0xEI0GgSSOAmB4YPWaCmos5Xw+8PMu2FhMMIE8kuU2kCmgezkHNngY0wzsEYwd/MH488RZ5orscIwgIAp4jIAaQ59jJkYKAIGARApq3B5maXeXuOVB5mFZyMDOWre/ab25KzF7NVpxccghPcY3onkxdUk8mxrSv5+w7YniQkwtGT1RUlEfJ+Zy1L/sEAUHAtwiIAeRbfC1tHQ+Gyx/6hCkSSjn7cRhNZMbka6cPVdQN2XlFVM5ZgxNioqgnM3eDJ2fFlr20hpnksWIGmZlB5ZDNmaH/+dpPnLn4KLv4m9OVzAl19phe9N4Pa+i7JTmELMOdmVj1kb9MoIqDR+mqp2dSOWf8DWO3/oRBGfSPP42l9dv30btMh1BRdYjSmR/pbzOGUxwznC/ckEcbd+xXmW+R2n0AM9mj7iPvLyBQXMRy5tObzj+F6RvSDXFBVub3f1jHb/I1nD05jM4c0ZX5zZIpv7Cc5q3ZqTIaZzAFxsTBnKmaMzzf8fJsytldqrLenjYkU6X+N2yYC1+ZuZwWMY9VHQfFDu7RnqktTiFkZM3OLeLpkhpFgYFs16DCyNldwueSz9mWa6lrhwQ6lTNjO3qjR8bpOSt3Mu1FDaUlxdJkpiZBZmRwZm3gzNqYDgElxKBu7RW2yzfl0QHOMg28UmLCVAZrtAGeIATodmgbo7jfjnCw70/Ld6jM3NBp/IDOlM5cXI4Ex+8uqlBZsbvw+KXEt2YOpRpF14EM2Mi2fTqTuLbmLLp7iitoVwEn+uPfBcYPtBtWCOhN1u4oImpeSGF1R6g7k1nidwp+rTzOMA4swFsHChdN4O3RYnugjyMprzpCq3acMHq2c3yPJwLqgc7J0dQ/I55G905V2JfzWCCjOehC4jl5YVKbE0lGjdrXpraQLBCrtsDzF4wCeonN/LvXfofgENRn8w7GcxKdBQFPEJBM0J6g5sYxVmWCPvOO9xX9gZmuwcOEh4mez8rMcY1Z52E2sE7p00l1iaR1f33yazbsyrxW4ZTeafTwlZNs7dz6wg+0io1AI0ESO/0jtzk/oMPCmjVYKYTyKcO70q1/HGVrBkbZnWxIgo/MXrg641+/FO3iH4wuTVpwWRzTbxQzfYFe8DDS85lp+0Cm+sDlE6gHP7A0AefXqzNXNNDDURtG5TBKrjl7qMccYeVsCL/29Ur6YVlOvd8c+gKNx4GD9Q2FHh0T6NJJPSk5pgXV1p7gE9POR/9ZxkYPlq2v3F7M7Ouc40e/043ttrFMOVBXS9Et6ph64MSBMIYimT6l6jdePa25BOakG9s/g2BEQmD4YkoLnh4YPshSDtEMN/UlSP7Db+rtWavpq1+z6/GG4WXprNHd6fKpAyk1OSmguMCshBaeOmTUhriTPsFKHXzdlmSCdg9hMYDcw8vt2lYYQKfe+Lbb/QbDAbecDwbzLDr7rg/q3ZC91T0tOYbevfNc+sP9n1BRWX3jwpu28VB8/fazFZnhDc/NYqoHTx/JnmsBo+3Bv5xKo/p0VOS4b3632vPG7I78+7nD6ZwxPe1KnX+Fd+lGxqKAedpcCfOCEofbUCte/a2MSvaWwVupFyQoXM2enlVs+Hjq6UF7aYmtCIHMBfuLqeI3Ulp9P862a9lGnTyiJ/1udB9l/MDzA8EKrrZtT+gbbAYQvHM3P/89bWOPqSPJZG/nG3eeTwltWgcMGaojXT0pFwPIE9QC55jExEQbHQx4FDWPsfZS4ommYgB5gpobx3hrAE37x38V0aYbXQZV1Q5JMQ28F1acALwM2XmOb/ae9jGVPUFzV+9kD415qgVP+3J0HLwX91wyju5/e56jKh6V4zH/7PVTqU8X8+ztf3/mWzYIedrLgbADiNowm0MM/7HDq56AyPWiyf2Y8ra58vSs2lFKuR7G9KDhtrGRNDQriYZ1S6LU+FY0kz0dOTy95UpgxtbwSnl2OKk/bWif/tsUxdatHR/MBtCD/FuZx4zxruTUwV3pmZumiwHkCqgA3S8eIPcGRmKA3MOr0WuDZTyUxWgKyYrz9YXxA71mLc2xQj2v2sDU5v99vMirNowOhiHwFk+R/N91U4x2NyhbsjHf0PjBNFMsGzz4i3CwAhypeSqP1NG/P99ABw95bkzGs1tpSNck9dcpubVNx0L2SDkzfpAPsYovLRg91fxp5Mx787tVbBCeYWszWDe27y01Zfzg/H5emUObdxVS29YOBi5YQRC9BQEDBMQAMgAlUIruef2nQFFF9AgwBA7axa5Ypd7anEIVbI7Aa1eCQHFNMEsUw4fA04MpLjtnj4qJgpelkg0O/J1MyOy+8dOGOxj8m9HTOaW14eore+MHxh2MLs3Lc9hEt+s5oB/xTW2irWF217Bq7M+F6/Pc6nIuG0Hnj+vu1jFSWRAIRgTEAArgUducWxzA2olqoYgAVkPtP1ClVsO5Or+CkoPK2IGnB/E9PKNVT+BVUQYHe1hg9Bh5Weod4OQLjJ6BmYk0ODOJMzPHqBgiJ9WV4YJYHnh3nHl5nLWBfYWlVUFvAMEb5o7sKXKfJ82d9qWuIBAoCIgBFCgjYaBHREsZHgNYpMjHCESEO5/+QAAwkhS2jTxClbH1lYFnB8aOZnTA8+KpxEWHcyAzGz2cryeTc/VowciO2tOWqWPVVl14Ie084P0LRKQLLBzpEkjl4W7eRyLDhaMskMZPdPEdAvKE9R22Xrd88en96PEPrY/18FoxaSBkEYjhqa/UhJOxNNqJIgcVjB4kKgQhKSSVl+Xv4JxJmE7SjB4zU0tam0afSexOgtGDP0fTW/rjwKiOJeowevTJCLM6JhMt2a6v6vZ2q8iW1J5zOwW7dON8TO5Iz87mg+DdaVfqCgKBhoAYQIE2Ijp9pgzvJgaQDo9A2Izmh2IgBKb35uXjzlZfeYrV6bw0PYxz30DAv6UZPfp8PYdqj1E2M6xv2ltNOzh1ky61kUfdtouPogFs8Azi89m+GQAAQABJREFUKa70JMeJCNE4VmJpeXnw6YhnC7l8XvpqhUpm6ZFSfNBkTq6JnFrBLmP6dqSXvlyuYrtcnUt0ZDidNjSLag9Vuaoq+wWBoEcg+K/uoB8C5yfwu5HdnFcI4r0tOGjkwkl9fXIGz/19qk/avY8TEfZ1Y5m4L5ToyBmh72c9kPXbSmnHnp8/ntqLSkpKKDc3l3bt2qW2YfwUltXQnLV76ZmZG+mWN5bRy99voeU5JR4bP5H86tUpoSXd84f+dN8FA+ns4R0NjR8kIoSHJzk5mTIyOEFhly6UmpqqEto5Mn6ASWx0BCd3HOIxPMnx0XTp1AEeHx9IByL7NzLGm5GbLxhDbTg5p4gg0BQQEA9QgI/yzeePVHQP2XnexzME0qkiYHb2U5cqlQo56HbOyh2WqffAFeOpT2YK3cCJ/Z79fKkl7WKV081M4wGKj0FZqXTd09/SlvwSS9p2p5E0psp4+ZZpirrgiWsn012vzaF9bga52vcH0vOs1NZ0xZTeVFy4V+0+wl6eLXvKaUNeGW3MO+Axu7rWF+Kjozi0BMHS+GvPnp7po3s2CLZGHA+msuDdwR+muFzF/mh92H/+bmR3qmY6k1dmrrQlTbOvY/Q9jak6HvnrpKAPftaf25ThWYoy58Wvlhsm8ATGV7PBOH1sH/1hsi0IhDQCkgjRx8PrbSJETb3FG3LpnjfmNlhJE8V0A+MHZtDa7YV0kPm24BXoxw/oMua3Wr1tH0/XcEQqC1zbCOgsLq/RmvToEw+yZmy96LMgh/MTtHO7OMrfX2GbcoiJiiAQTu5j48ZehnRvR49fc3q94nU5++ih9+ZTCdNCaIGzEdwustMi23AJnw94vFpyWYeEGMrjlSr2VB/g1/ro3hm2bKHooLy6mi5/dCaV2dExYCprPE+TbGROpJKKaoK+fbokq+mfpZt3q/rINIqA4F4ZyXTbBaMUv5Ze6VlLt6lszFg1BV1atmhOnVLilPdhCxusePhCsKR8IPOixbWOYH6wYnWOcTxOvTolqvqrmapju+ICO6Z4svqxh2ndjkKVx6aWDRHgncgUDeeN60W/n1D/AQV6g68XbqH5a3ephJKYssnkjNUdWY8lm/KpoKRS0W+gHA/24b06KIN6z/4DFBl2nDq2bUW90uOVV6uw/BBtgsGTf4C27eFYH2+WbfF5h7doRumJURTFFBTNjtcqmpHEmFZM5ZFEmMLT+NVg8OAPnp7IyEiPDR4FtsF/ufvK6PP5m5gXb59aHQb6kQFdUxQ/2qqtBUzBckD9ntOSY2ncgAyC4QR6CHsJ5kSI2rmAl+0LxgK/ubLfOOn6d02lc8f2UpxwoPsAyWtxcWi9cOH8JRO09isIzk/JBB2E42aVAaSdemxsrHpQ4DumKhCnEWqC1Obl5eVUzcZLqAmmcjB1A+MK6dwbU0DeiZieyspKxcGFWKbNu8uV0bOJY3rAveWtpHI8z5BuqdSjfTR1SeHl6nD16QSeBnh14N3RDB/NENJVC8jNUDCAXAErBpArhAJ7v2SCdm98ZArMPbyktiAQNAjAyIIRCYMHf4eP1Cp+rc1s7MDwyWPaCc3b5ulJwcDJahdL/ZhhvV9GAntVWisDHavF0L9m8GjGDgyfYDF4PMVEjhMEBIHgQEAMoOAYJ9FSEDCFwLFjx5TRo3l6dhdXUTbH8mDV1ta9FXTkZApmU+0ZVYrlpIR9OvK0Wad46pkep6ZWtXoweDCVBW8JprNg+IjBo6Ejn4KAIBBICIgBFEijIboIAh4ggGnQqqoq5eXZvZ+9O2zsZLOHB0HMB2tOxCF50KztELZp1HRWbzZ6+nSKU6u1YOhA8KkZOvDuYAolKSmJCgsLOa6GsyKKCAKCgCAQoAiIAeTjgUG8B4LvrBIEKGqCt+vwcF5SE4KCOJFQ9BxohgOGzNPfBaaWNIMnb28RB0zvZ6PnACGOp6TikCW/hjachblPpwSe1kqk3uzpQdA4BGMCQwdeHvz+sK0/J3h+INgPPUNJ9L9HnKen4xfImOD+gvMMxXPT3ytxf3GWRiGQx8iVbjjPUBw//Xjh/mKFyCowK1B00gYeVHhQWCV42GgPnFB9w8YNOFTPTRs/GAfuGAjIxYO4mtw9hbR8M1jYizl4uZQKSq0JFG8R1oy6p3EcT+ck9dcxOUb9ZHEzxc1G+zPzWw7l8dOMIHfHz6rr39ftaPcWd36bvtbJyva18Qvl+wvwCsXx0+6dOD/9+GljinJ3RQwgdxFzs76sAnMTMK4uq8BO3MDAubV7XwkvWd6rjJ2tvDQdCQmtEmRgRgwP/rq3jyWkM8CbsTalBWNH73E00y+OT0hICMkpMFkFZuYXELh14BWJiTlh2JeVlRGur1ATWQXm3ojKFJh7eEltQcBnCMDLs3NPEa3ZupfW79xP2zhoeT/n5rFKYjgTYY+0Npz3J059IrcQDBZMY8HowZ/ezWxVv9KOICAICAKBiIAYQIE4KqJTk0AAK7ayd+5THp6NbPBs3VtOB8AqapFEtGxOWe3bUE82emD4ZKTyiq3fVmbhE8aPNuVhUZfSjCAgCAgCQYOAGEBBM1SiaLAjcIjz8Cxcs42Wb8rlrMQ8rcVxPNWHTzCrW3Fu4Fbr0i6GuneAwcNeno6J1Pq3YGUYPFqAshV9SRuCgCAgCAQ7AmIABdEIPvHhApq1dLtNYyTZfeGGyfTFgm20aONuquUcL5FMjTGNCVQTYiPoP1+ssNXFxlXTBvLxOZwA76CtHHQVd118Ct31+gJbGTZ68cPzOKfJy+ZAW728fvvZ9NznS2jzriKm5ahTcSPnjOnJFBwR9MKXKzj47kRtJMh7+C8T6NnPljJX1Uk6jKTYSHrquil0x6s/URFTSKB6DJM13vbHkTRzYTYt3XyCiwqttIpoSe/cPoWue34O1z1BkRHGwbrnjOpOY5jG4jnm+QKPWAsOmu6ankB3XjSWPp67nuas2EGHjhxTK5fOHt1DUV48+sECymUaAI6uoQ6JsXQrU1tU1hyhbxZtVVQbMa3CCQziI/um0atfraS1OYV0+OhRase0G1ecMVDVfZnZxYuZNqMl00r07pxMt/9pJLVmw8KRlHJdGDrrcgpoc24x7dhX0YC+w9GxZsqxEj0tsZVKRHjs2FE6xrTszVu0pOTEeBo/vK/iCFu0pZDKmfIgnqk3BnZjOg4mCf3sl008rsWK/qFL+wQ6f0IvquVjV27Zy1Qp1XxOvAKMaUG6pScZqlFedYiWZ++lguKD6veWlZZAfZl7bQfTeWzOK6GjdWGKZgOcaW2YngTt5u0vJw7fV3QLg1mPCP6d+kJA/QLd9hRVMK1JM6YFSaABrMde1nUN060Ai7iYSKbCaMd0IW18oYJP2zxwsIbPbw8V8jUVydduj45JTKnSlqlLSgleRJx/UptoGtqjvaK/WbFlj6KoARad28Wr34ARzYcVSoOKB7oB65ZMJYLfRf/MVJ5Wbci5fZgpXFYyDQloSng5gKKEGcw8e7h/iQgCjYWABEH7GGkrgqD3ssFy0SOf+1hTad4TBEb2TqOHrpzEhs1x2skGwFo2djbs2KcMgf0WBixruoFFHX94psDYhAOphv+8pO3Smq/32ZMfrLeyYYoHJwQG1luzVtOn8zYqY1tfGQ/VI0eP6YvUNjjIjvJxemnDRthVZw0hEHRaJVj18vHPG+i9H9ZSDT9c9QKONrwc2MuIXml08x9GssHQyn6X4Xd/BkED29e/Xkn/W7C5gRHtCHsYPfZ8eYj7uvacYTRhYGfDc/SECgO8gO/NXksfzllPINHVS2pCa7r+vBEErDUBh96rM1cwL9thrUh9gjPvyt8Npmn8guMLkSBoX6DaeG0KF1jjYW1ZT94aQMjoe/Y9YvxYNiAWN8QOKYriN/HaY3V02O7m721X3DRlMs1ETGSYIuzEM/wwP18s7sapmiDbfepvU/htPpHJeOfQYvY0WiWXTRlAl/CfFfLEh78q76a7bcH4eeGmM6ltnOu8Iv4ygGBA3v7SbOXBcvf8HNW/dvpQOm987wa73TWAYHg++M4v9MuaXQ3a0grwO779wtF0+rCu9DYb0O+ykepM/jixD/112hBnVTzaJwaQR7AFzEG+MIAa+iYD5nRFESAgxk9g/Q5g8PCLtU3Y7qHKQ0ctMX7QLLw7nINQ/fGqdO4sjNbsPEBlvBgM8dGNafzgJOFNuf+tufThT+ssNX7Q9tvfr2GPmfeEsLOX53hk/EAHTPk9/N58bAasvPvDGkuNH5zoS18up635JV6f81e/Zjs1ftABprn/7+NF9NOK7S6NH9T/aM4GWrbZOkMbbYoIAkYIiAFkhEqAlMH7I+I/BGCQ6I0daAKDx6rpJrTNzh2KYqOHnUjqQcG2FIGUHX8cxkQ5ew6ocv+hQCrO6r8/rfeJCv/9cZ3X7XrbxrrthTxtWei1Hr5o4BAboJ/O3Wh50zBKPmCj1huB98cs9vBivcpTeGbl/dne6Wa2H6nXtBEQAyiAx//CR2YGsHahp5rOsaNODg8Jq4wdNIi4nXA2ePCHvtD2ITZyEMNj4WIwdGW54EHsC1m9rYDjc+rHjbjTz77SShXk684xRnWXcfBuIMp6NsysnlrVzhMBy97IjoIDTL1iPpkgvG1mZeOu/VR9yHseO7P9Sb2miYAYQAE87tUWEFkG8OkFnGoweKyScA68xYo3/lDGDtpFLDC8Ovizsi+rdPZHOwjSLeOVWZ6KOw9VZ30Ul5l/ODtrx+p9Vp2fkV41bHV7Y2SUuGHQGPXvrAwB/qW84k1EEPAlAux4FwlUBLCCBsuTRQIfAUxnje3XkZePp/BfKi+/jlcrpj7w0dRR4CNiXkOs/vFUvDlW36dV7ejbtGLbl3ohVYU3y86jOV2CL8WX5+5LvaXt4EFAPEABPFZ/n2H9SogAPt2gUs0+NggZle+9/FSaMa43589JVLlPkCsoVKQ5kg75QLrwEvsozvfkqaQn8yo5ziPlrSDvUSBK74y2PlMLbcMI8lS6dkigCBWpb64F5PUyK+2TYiiOc0iJCAK+REAMIF+i62XbvxvVy8sW5HArEDB6RNjHBp02NLNBV0N7dCDcyL0RJCX0t4AkddyATj5RY/qYHl61G8ZJMJH40xtJjo+mEb3TvWnCZ8cmxLbiBJ0+wn50T6/0RjLLqW7kcpoyvKvp/qZzAlMRQcDXCIgB5GuEvWx/2sgsL1uQw71FwFW8Dt6Cb79gdINuMIWJciSk81ReuPlMGtazg6eHW3Lc3zhx3vXnjqAUNhSslCGc+XfqCO9/3386rZ+acvREN3hAbues4L7KjuyJTvbHXMf4J3AmbytlDE/Xjh+Y4XWTyJKe1jbWZTtn8Dhfw7mHkAXclfTjrOLTObu8iCDgawTC7mfxdSdNuf1Dhw7RUaZU8FRO6d2RVmbvpv0BGqTp6XmFynGgEpn5yAU8lWD8LpHCmXB7ZyQTVhkh/b9ZwZTTk387nbozHcWkIZm0jXO25DO9Q2MKPD83cBbf343srmJFxrAnAsvF3Vn540hfZCK++5JxlhgeoF0YxzQmyGuDVWFmJZanzu67bDzTRpgzMDHG0cytBsE1jWu7MQSxNiP7duRcQAVeBYxruk5lT8xtnJgQBrq9gDMuLCyMqqvNBYWHt2zBtDSdmNduP1PKNDwGpv+543rR388dzi8CzWlMv060t+Qg7WIKDCMZzed532UTfEKVEh4ergiA0a+392Uj3QOhDMkea2tr6cgR60iVA+G8oEOrVq3UbxPblZUnr/OYGM+97EKFATR9KN5mgtZUc5QRGt4H3GQO6TLkgdeqsvqIT1Ya8QpulRxHn2I/irP3HeYVJfbTQlgBZcBAQJGss15f7Rwb6xNv/QO6plBibDQzsRco/qQIvpF3Ze6i7hy/8z3zpWFlElaiAN++HB9Syst9d/JNWztvZH/+w4Q+pjMZ1xyupR+W5dDG3FI2IKoVN1ePtHjV99cLt1AF8yihP/A7gTYAxoG9YEn0m9+tVlmhQe0Qw8HDw3t1UDEw89fuogMHD6mYDtAPnD2qB1Vxn0g0qOcCw2/lm8VbaT942LjDRJ5imcwZehFvgWXRWHXUOrKlCuaewuWJdjQROGbBujxatCGPCvhBBs8JOMNG9O5A63fsp41M8FrDy9zatGqp3vZhAP7K9U9wgZHiAps4uAv18VF8FBLozV+bS7sVF1hz5Rka1SddcWWd5AJjXrSuqXQ6GwLuBPL6KxO09jsA3cr8NbnM+5fPXGCVNi4w8H6t2VbI2IMLrJba8pjBawjOs4Xr8zlNQLkyPjq3i6PT2JjuzvxhjsTdTNBaO/hdIEv4gnW5zAVW8RsXWCJN5qlhjUpFq4tPGNJzVu38jQuMfxfMywajuD+Pi69EMkH7CtnGadcXmaDFAPLx2FllAGlqxsbG2t5CS0pKAs7Sx4M+j0lH8YZn++N8ISAtbQxRBKFt21A3Nmay2Jjp2iFRkTJaEShrhf7JycnqLQYPjH37vM+CbIVOVrYRERFBCQkJVFhYyGSrobWC0d8GkJXj5KgtTw0gR+0FUrkYQIE0Gu7r4gsDSJbBuz8OcgQjANbpvH1s6BSWqbc4MK2D2dmdKQhvgQTJJVYRdWWeKqxIwV8XXn7uzaoib3WS4wUBQUAQEASCAwExgIJjnPymJZKdwbiBGx0GjrbtywRtRicbz1M0mezVyWzPf78ZO1gCjbgCEUFAEBAEBAFBwF0ExAByF7EQrH+k9ijtKT6ojBxMX+Xvr+BprDL1Wc1TWo0pLTkwMzMtScUNdEqJYY8OjJ54FTPTmHpIX4KAICAICAKhjYAYQKE9vrazO84RyvsPVNJuNnR2szcHK4rg1YGxg3IOSWl0QbAupqwQJImprM68nc5LatPSOlB5ebnplSiNrrh0KAgIAoKAIBD0CIgBFPRDePIEEFiLqak9bNxgFQy8OuoT2/znL1qNxNgoykiNU4ZOBgwdXo2C7xKrc3LsZEsQEAQEAUGgcREQA6hx8bakt1JeRr0ldz9t2p7PgcgHeNnpQTZ2KtSnr5ijzSgOQ6cTGzYwbvR/rS2gKjDTv9QRBAQBQUAQEATMIiAGkFmkAqjeZ3NW09MfzfeLRlhmnhLfWuXt6JgS95vB04awLeSFfhkS6VQQEAQEAUHAAwTEAPIANH8f0j7Jdep5b3VEFmDE46RzgrKOyTBwThg5KAMHkIggIAgIAoKAIBDMCMiTLAhHL529LVZJ27hWlM4GjvqDwYNtNnbA+wSGcxFBQBAQBAQBQSAUERADKIhG9dQb31bagmLCKgF/TyXz/qzaWmCqybaxkVRU0ZADqUNSa45DOsnPgsZ6dIynnN1lTIdxcokZaCh6dkzgtP0l9fpLbhNFZVVH6MjRY/XKh2Sl0opt9TMmRzP1RrvEWMrZU2qrC1sNHEc5u0tPUFYcq1NcRz06JVFfpl34/JfNdJiX+0Pg3TpzRFdav6uIduw5oOgtWC0C8/Z543rSf39aT5U1TCXCarcIa0bDe6YRchovZ5qFo9wu+sJ03/TR3emTuZu43ZM6J/B5XDyxD302P5tKD9ao/tq2iabLpgxgHqIWlP39ekWr0YZpCrLaxXIgeAt69euVKoEkVuq14XxHM5gIspBX5v2yZpdKONmCuZkyOHD8lvNHqilH1ajuP9AP/G/+ZkWPEMb6YlXdBRP7Mn9cFVNhFFJ51SFFdTGoW3tFneFt7qSt+cWKYgGcTqAKyeJElOMHZCjddWo53MR5grICvzlgBE6ufkyBAMoK8HrZy8xfs2n2iu2EnFSgLMGYXs4knJiKNSPHjh1X9BFr+HcELIDxoKx2TN+RZphHCnqBFqSI8QNVRu/ObRWHVUyLxrtdHmBcMP74jR/h31e7pBgCTxawFhEEggmBncwEsIDpabAoB/dTpDbB/cKeZscf5yRUGD5G3QoqjPve+IkWrN9t0xQchmAZOGlW2HbJRogjMIyJOx+7+jR1luAru+7pb5iTq77h6QwCcC7dxTxjyJrtrlQyX9kTHy1UXGD2x0bxtOhVZw+h35/a3ykVBuhRHnrnF9rBN0V7gdfxnxeNJbCBQ1Dnpv/MooPMa2ckZ43uQTcyWaszyc4rpoff/UXdfO3rwdt59yVjbUYFspij7kY2jO0lng3W2y8cQ9NPHaJ21dTUUFmZMaGn/bHufv9i/iZ6jY1ivWGttYEHxy1/HOkWh5l2rJlPocIwg1Lg1klJSaGqqqp6ZKH+0vYQkz8/8+limr18ewMVwCF4Kb8UXjCpb4N9jgp8QYURJmzwjuC2ptxb1uE3v1lB3yzJqaeMP3L21FNAvvgNAbxFgcB1PBNHXvDAp+xBaeiNc6ZcedVh+om9Kaf0TncruSQ43m58/ntmJK/vjdP6OspeliWbdiuP0PC+XdRNGGkZ9ILkmtc9863yTOnLtW0Qef60cocygOC1+cu/v6JDTKzqSLawcYMVkGAYNxIYPzc+N8shg3oFY/Ej35xBPguS2+ue/lbx2Bm1hZv5z6t2sLHUViXq9BUb/Lvfr1EeQY10114XGJCrtuxVpKZGbO729d397i4bvLvt+7O+sME3Hvq4H9z+8mwmTc437BS/b3hacV8Zwi91ZsQXbPAWTqaYOQWp4y4C7/+0wd1DpH6II7BueyFd/+y3ivnbk1Ot4Yf5o+/PZy9ifQPFWVtvz1pD23VTjo7qvvzVMtqa19CDAmPo0fcXUDUbOc4EN85HWDfcPB0ZAfrjf2RjDizo9gID6pH35jeYUrWvBy/LI6zXU58sVlOT9vv132HP3fXyt3TgYLW+2LLtrfkl9DYbQK5kC9d7b/ZaV9VkvyDgNwQ+mbuBcJ9yJZ/M3cjT9MYvVa6OtWK/GEBWoOijNmYvy/ZRy9JssCOQw7FL3sj2vQdoFXuSzMhhNphmLjT3W4SR8O53yxo0u37HfsKD24wgLs2dab0X/9ewv8Ub81USUDP9wauylL1XZgSxYV/MXWemqtt1Pv9lk+ljvlyQTbV28XKmD5aKgoAPEcCL1WfzzP+WP5230YfaOG9aDCDn+Ph172MfLPFr/9J5aCOwxqQBhKkko3gUR+gs25TXYNeaHHPGVoMDTRTsLGgYi+Noqs5Ecy6rGJ2fy4NMVHAHI3D0bdttzqA00bVUEQQsQyCPKZYQn2hWxANkFimpJwgIApYhgBgYM4KVU+7IgYoTq9/0xyD2yFdi5Akpd+MG7K5eyMTuC3HnoYH+fYmpL85P2mwaCLh77SH2D1PW/hDxAPkDdelTEAgABOJ4ZZMZiWsdZaaarY7R8ta4aHN92RpxYwMrSuwlLsY9ne2Pd/Y9kdMl+ELi3cQ5jpfziwgCgYaAu79LpBQJw9JmP4h/evXDiQZjlw9eMToY1RadgwSBgZwLx4z06JhIkW5k/x7ep+GqrIHdzPVlRh/7Ol3ax9sXkdlza3CgiQKj8zNxmMsq7mAUHdnStnzfZcNSQRBoRATAGgBeSLPiy2vVlQ5iALlCyI/7R/fr6sfepetARqAnJwP0RrpxQr0BnHzQjIRz8sFzODmjGUGiy0vPGNagah9ORtk7o22DcqMC5ANK46zkZuVvM4Y3qDq8VwfFV9dgh0FBJhtQSDJoRmLZk3XuhP5mqrpd5/fje5PZ3OvnjuulEn263YkcIAj4GAEwCPx+Qm/TvZzvRl3TjZqsKAaQSaD8Ve2Gc04kXvNX/9Jv4CEwtHt7evq6KRTTKtwj5eA9uPPiMW5RnVw6pb/KwOyqw+vPPYW6dDDOVvyPi8a41Lklpzm/59Jx9MS1k0094KeN7EbdOzY0BpHt+q6Lx7r0XEVFAIuxdOP5p1AyG17OpDnf2B+77ncEI8gXAk/WldMGu2waxuSfTuvnsp5UEAT8hQAM9CF8n3IlF0/uz5nWk11V89l+SYToM2hPNOxtIsQeGcl0oKKKlxCfpH3wscrSfAAjMLZ/J3r4yklqznzaqO60aH2eW8GwWWkJ9NhVpzG5rXt8cpijP3VQZ6boqCKktrcXzOPfxEbEjPF9KCoqyjARIigvkLRwc24RFTOthb2kJ8fSQ3+ZSL34N4/20N+8NTs5GeIJChN9fXhK/nhqHzLy/mj1QG0ynJMcbthZaLgqJZOzYT/610mKOgSGEJJLgnoCGaHtJTkumh7486k0+ZQTb7a+SoTYp0sKJbVpxTlU9vEy94aBoZOHZrKBOF5RgtjraMV3SYRoBYr+a6N169ZUW1tLR44YZ09vLM3wsjBuYAaVVx6mbQbpLzClfvVZQ+hCNwx5XyRCFCoMH/8irKDC0FScfPPbzKulfTvx2bFtNOUVVdUv5G+tIsKo+nD9LLoRLZvzcma7Brhuu9gIKqiov0oHrsHmzNsC7iu99EiPo+z8hsuOe3KcyOa8+styR/ZKpZXb9tfrsyU/SMdwFuKfmb9KL13at6H9pcxLZpco76xRXenrhTn1aD8SOcC1U2osZ0QutJWH8QV30el9VRkSytUeO0YIju2fmUpDu6fSm7PWqqyjOJuo8JYqBftKzqi7Yed+Trh3nHDBtktozZxd/enFr1bwA/Mwt12n2sCD+Dhvz121S+VeacYTFXGtI+jPzEf1EtfV65zOUzd/4bf4t79bTUX8kMeDOjWxNf2Vy+Aazt5drhLuxfM5ZLaLodiolvTClys4Z0051fHQxDPX2h+Zxws8VN8vzWEaiMPKE4Ipq5v/MJJSWEd7wXkghwyyIoMLLJO5dv40uR/tY4oMLAev4FVc6A9z7YM4FsdbklvkzVnIhhcMBfCqQbdRPIUEoyUiIsIpFQZ0R1JE6AW9wXkFjwroL0DzYRQMOYezQ3+7ZCuVcH6gCL5xwgNy8en9TWeyRl6SFVv2qD6xQgVB3Yi3AR8YpuzsZRMnVly6eQ8Vc3/RPD7obwT/ZltFRlDbtiem8XxJhQF9QDuygDEGtx348donxjDG6W4brvbn5uq7UGG4Qiiw9wcSFYaG1J6iCvqVf8v4RPZyeDrxIgROPnfEF1QYYgC5MwIe1LXSAEL3sbGxFB19wlVfUlLid0vfA0hcHtKuXTsqLy+naiZpDTVJTk7mh3yYMgL27dsXaqdnygAK1pNuwWSojWUA+QsjMYD8hbw1/QaiAWTNmRH5wgCSGCCrRkfaEQQEAUFAEBAEBIGgQUAMoKAZKlFUEBAEBAFBQBAQBKxCQAwgq5CUdgQBQUAQEAQEAUEgaBAQAyhohkoUFQQEAUFAEBAEBAGrEBADyCokpR1BQBAQBAQBQUAQCBoExAAKmqESRQUBQUAQEAQEAUHAKgTEALIKSWlHEBAEBAFBQBAQBIIGgRZBo6kPFT18+DB99tlntGLFCkLenqysLBowYABNmTJF5WzxYdfStCAgCAgCgoAgIAj4AYEmbwCVlZXRtddeS/n5+Qr+hIQE+v7779XfokWL6L777qPwcM84l/wwntKlICAICAKCgCAgCJhAoMkbQP/617+U8TN8+HC65557CJlQ9+zZQ3fddRfNnz+fnnvuObr11ltNQOnbKl8vXUdPf7jKt500UusgH6hPsMG0G1wYFdGCqg6d5H0K48K2zIu0j/mn9MJsD0xnEEklBw/ZinF827go5qqqsZVhA32hvKi8hrMvn9iFstSk1pxF+5iipdB0acV8UBOYvyY7r5gKmEYCFBkxzF91Su80um76MDaE618u63cU0mvfrKRdBWWKrgDH9+2SrHiuQMdxjCkYIKCnGM40Dw8xh5c3kr+/nL5bsk1xadUcruXziqbhPTswXUM4LdqQz3QaFQSqkS5MhTGyTxptYxqFE1QYhykuJlJRP3RMbkOLNuYzn1eZykadzt9H9kmnUqakWA76B6bvAKVFHz6PM0Z0o1QD6g2jc9jPYwS6io27iphy5BjFtmqpyBCnDs9iComWRod4VQYajTeZbiSvsIxpT44TCF5B9XHjeSOoNY+ZiCAgCAgCrhBo0lQYmzZtoquuukqRN86cOZNABKhJaWkpzZgxQ02BffnllxQTE6PtcuvTCiqMU298260+pbL1CIClHCSkGsPxfW/+TAvW5bnVEVOB0WePXEa9urRzmwrjvR/W0js/rCHwWjWWgLfnKiYsBLOzM/nq12x68ctlhuSdMFTvZXb3AWycWCW3vzybVmTvNWwOnG63XzCKJg/rarjfm0KhwvAGPf8fC6JQ7T4Ozz/43EJNhArDvRFt0kHQ8+bNU2iNGzeunvGDQkyFDRs2THFtzZo1S9Xzx39i/PgD9YZ9gpn7Dn7w5uwuoQffmee28YMW4YE6959vM3nqSc9Vw54alrz7/Rp6a9bqRjV+oMVR9qy88L9l9MX8TQ2V+q3km0Vb6NnPlhgaP6hSxuSjd7zyo/JaOWzEjR23v+TY+EEzxxnkxz74lRazR0xEEBAEBAFnCDRpA2jjxo0KG0x/GQkMIMi6deuMdktZE0MABsw/Xv2J5q3e5dWZT7/1TdPHY4rnHfb++FNemblCsdPb61BaUa0MJPty++8wHp/8aKHXBhymvVbwnxl5+P35ZqpJHUFAEGjCCNQPamhiQCDWBxIXF2d45lq5FiBtWIkLi4qK1JSG0f7jHEfSvLlnduY57C0QCSwESiu8d5vvKSpXJ2Xmd/Etx/zUacFLfoICBsz3y3Lo0ikD62nw48qddLj2WL0yR18Qc4T4oP5dUx1VcVmOmB+zUn2oVsVFje7XyewhLuvpx6sZT7Xpv7s8OMgqhOK5Ycw0CeXxC9Vz048ffp9W3BebtAFUVXUiuFYzdLSLQ/uMjY1Vm1o9rdz+c+LEiYSl9Eby2GOP0TnnnGO0y2XZgarjLutIheBEAEZQWkqKS+W37S5zWacxKuTsLSfEF+hl+95f9V9dbucWV9PkUfXbcHmQrkL+/grdN9eby7cW0rmnnfDiuq7tXg3EC+pjBt07OvBr24914GvsnoZY7IK/UBTEOuEvlCU5OdmS0/PMNWFJ1/5tBJ6ZQ7/FYmiBcfYaaT8iR8aNfX35LgiYRWBL7n5TVQ9WuxcvZKpRDypV1jQ08I3KnDVdWd2wDWf17ffVHjXnbdKOq6gKDOw0feRTEBAEAguBJusBggstKipKrQRwZOBo5a7yAA0dOlQFSxsNLbxLWjtG+6WsaSIwsk9nU7+LZF6Gvn1Pid9BwpJ7+98xytyRJE5pYN+GO8dHR7XkKbeTaRJcHdsuMcar/uzbhwteuxccO3aMjh41r4t9W4H6HSvdcJ61tbWBqqLHeoWFhRHOD4Lzw0twqAl+n/ht4i/UpGXLlrZpZ/19JCLC87QXTdYAwo8jKSlJ5QA6ePCg4W9FK4+Odn6jf+ONNwyPRyGWwWNJvSfyzBWD6MY3QyP3jyfnH6rHIBQhMrIF7du3z+UpDuqaQovX57qs5+sKAzOTG/yOB3RpS1+7MQvWu2Ncgzbc0btfZopbAehTh3b2qj973fTL4I8cOUJYSh1qgmkhPGg8vWcFMh76ZfAIawjVZfA4r8rKykAeCo90S0xMtL2A4LmqxQC1a+d5io0mOwWGEYABBNEMHfVF919FxYmYg/j4eF1p423269ev8TqTnkwhMLxXB2qBrIteyIzx5sf1jBFZlBAT5UVv3h/aPimGTh3UuUFDYwd0ok4p5uIoThuSSfDIeCNIcmgW+q4dEqhD21hvupNjBQFBIMQRaNIGkBZItWPHDsNh1sp79uxpuL8xCn9+5rLG6Eb6MIFAYmwr+tcVE+nuS8ebqG1cJYmnjR66+kzjnQalUZxd+v4rJlBEyzCDvb4vQobl+y+fQC1bNOw/jKeR7+N9Ma2cU8V0aRdP15833GtlY6Mj6ebzR7psB5g9dd3pLutJBUFAEGjaCDRpAwirtyA//fRTg18B5od//vlnVQ5iVH8KpsJE/IvAgKxU+vDe8ziGoDmN7d+JnrhmsqKMcEer7umJNO/F69w5RNXt0zmZnr/pTOrZqW2DY5H5GPQX9hIZ3tBgQR0jQyqczwnTcvYygJesv3TLNII3xZFkpMbRSzdPo8HdGrqhoRs8WM/dcAZTVTg3khy1b19+xind6EE2CGHkGElm+3j6+L7zhA7DCBwpEwQEgXoINGkqDMwhXnLJJbRr1y668847aerUqTZw3nvvPXr11VepU6dOhG19DgJbJRMbVlBhaN3MuPNdKquuH7h30wWDGnCEwVlwzqju9Mn8Ldqh6vMMjolYuGE3ldecDHDEc++z+86mcx/4ql7dqcO60Pa9pbTVbhk2PFJn3PYeHdLlf+nTOYmmDO1KT36yxNYG2v3P9ZPoiY+XU25hua08NT6a/u/vk+maJ7+liuojqhwP4PsvH89tHqVXvlpJB7kc1BPDerSnf148jv729DcEbi2wQEQyX9g1Zw2mzA6JKuNvUVkVgTOsW1oSPXj5qfSv9+cyTUKBSroHDq7Th2bSpVMHEagrlB7cBqZ07rt8HJ/fAXr7uzVUxpxiaHfysEy6ZHJ/+mH5dm5jDx1mrrCOPMUzg6kgEMBrJMg4PG/NLtb5MCFg+SzGvTUH6z7+wULata9McZFlpSXSnReO5qWpkQSvI4Ix8dszEwNk32fOnlLKzi2i6t+4wAZ3a0/hPOCrtxXYuMAy2WDpnZHMvGiVtDZnH5VXMRcY9w2urMTYKFrHHGY7CzCHTpSeHKtoKqpqjtCqrQUnucDY6OrExo07ksfjnJ1fSseoOYU3O879pVACe818JfPX5tKv63KpknVvz9NrZ4/tQeltzU3JeaKTPgYIcRahHANUXFzsCUQBfYw+BkioMAJ6qAyV08cA4d5pRQxQkzaAgDIIT++9914VNT9hwgTKysqi9evX0+LFi1Uw4DPPPEPexOJYaQBBX+Qm0oKyS0pKHK4+Q91gFQS1lZeXU3V1dbCegkO9vTWAHDYcIDuwIgM0MoWFhSG3ykYMoAD5kXmohhhAHgIXIIf5wgBq6DsPkJNtLDXGjh1LTz/9NKWmptLcuXOV1wfGT0ZGBj355JNeGT+NdQ7SjyAgCAgCgoAgIAi4h0CTXgavQTVw4ED69NNPCR4V0F7gLR0GUSimg9fOWT4FAUFAEBAEBIGmjIAYQLrRh4sNfyKCgCAgCAgCgoAgENoINPkpsNAeXjk7QUAQEAQEAUFAEDBCQAwgI1SkTBAQBAQBQUAQEARCGgExgEJ6eOXkBAFBQBAQBAQBQcAIATGAjFCRMkFAEBAEBAFBQBAIaQTEAArp4ZWTEwQEAUFAEBAEBAEjBMQAMkJFygQBQUAQEAQEAUEgpBEQAyikh1dOThAQBAQBQUAQEASMEBADyAgVKRMEBAFBQBAQBASBkEZADKCQHl45OUFAEBAEBAFBQBAwQkAMICNUpEwQEAQEAUFAEBAEQhoBMYBCenjl5AQBQUAQEAQEAUHACIFmdSxGO6QsMBF4+OGH6cMPP1TKvfbaa3TKKacEpqKilSECp512GhUUFFB4eDitWrXKsI4UBiYCW7dupRkzZijlJk+eTE899VRgKipaGSLwyiuv0H/+8x+179FHH6Vp06YZ1pPCwETgwgsvpHXr1inlFi9eTDExMV4rKmSoXkPYuA0cO3aMamtrVadiuzYu9lb0hrHDX7NmzaxoTtpoRASOHz9uu/ZwHYoEFwIyfsE1XvbaavdOlFv17JMpMHuU5bsgIAgIAoKAICAIhDwCYgCF/BDLCQoCgoAgIAgIAoKAPQJiANkjIt8FAUFAEBAEBAFBIOQREAMo5IdYTlAQEAQEAUFAEBAE7BEQA8geEfkuCAgCgoAgIAgIAiGPgCyDD7Ih3rx5M+Xm5iqthw4dSomJiUF2Bk1b3V9++YVqamooLCyMsCReJHgQOHjwIC1cuFApnJqaSgMGDAge5UVT2r59O23btk0h0a9fP2rfvr2gEkQILFmyhMrKypTGEydOpJYtW3qtvRhAXkMoDQgCgoAgIAgIAoJAsCEgU2DBNmKiryAgCAgCgoAgIAh4jYAYQF5DKA0IAoKAICAICAKCQLAhIAZQsI2Y6CsICAKCgCAgCAgCXiMgBpDXEEoDgoAgIAgIAoKAIBBsCAgXWJCM2OHDh+mzzz6jFStW0IEDBygrK0utQpkyZYpaURQkpxHSapaUlNAnn3yiVpsUFhZScnIyde7cmf7whz9Q27ZtHZ57dnY2ffrpp2p1X3R0NPXt25dOPfVU6tKli8NjZIfvEVi9ejW98MILNHz4cLryyisNO5Tr0hAWvxTivjhz5kzasmUL7dmzh7BSb8SIEYr0tEUL40edXHt+GSrDTvFs+/7772nXrl3qmZaRkUHTp0+nnj17GtZHobfjJ6vAHEIbODuw9O/aa6+l/Px8pVRCQgKVlpaq7bFjx9J9992n2MUDR+Omp8m8efPokUcesS1xR3oCGEQgzYyKiqI77riDsHTTXmDUPvvss6q4devWdOTIEfWHYx577DEaNGiQ/SHyvREQwJL3yy67jPbv30+TJk1S15h9t3Jd2iPiv+8bN26ke+65h4qKihTRMK6/4uJipRAeoGCBj4iIqKegXHv14PDrl2eeeYY+//xzpQPGCWSnuBeCNPrPf/4zXXrppQ30s2T8uCORAEfg5ptvrhs9enTdLbfcUsc3XaXt7t276/hHocqfeOKJAD+D0FYPY8E5fdRYvPHGG3WHDh1SJ4zPV199VZXzQ7QuLy+vHhDr1q2rGzNmTB17e+o4P1Ads1XX8UVfx14kdcyECRPqCgoK6h0jXxoHAX6pUGOA6+7+++837FSuS0NYGr2wvLzcdv2xoVPHL4dKB/Yk1P3pT39S4/j000/X00uuvXpw+PXLzz//rMYI98GvvvqqrrKysq6ioqLu448/VvdHXIPLly+vp6NV4ycxQA3sysAq2LRpEy1btkx5ER566CFq06aNUrBDhw701FNPKVfhrFmzCG+sIv5B4Ouvv1aeH3gKrrjiCtubJt5kMHUyfvx4YmOIUE8v77zzjnrTueiiiwiePLztILnX73//ezrvvPOotraWvvzyS/0hst0ICMyePZvmzJlDcXFxDnuT69IhNI2+A54DJBfFVOV1111H8fHxSodOnTrR1VdfrbYxpvDGaiLXnoaE/z9//PFHpQTun2eddRYhDCAmJobOP/98NYWJnZga04tV4ycGkB7VANzG1Apk3LhxFBkZqba1/zAVNmzYMOUqhBEk4h8EECsCYW+OoQK4MUNycnJs+6urq5Vhi4LTTz/dVq5taGXffPMNHT16VCuWTx8jsG/fPvVigReMCy64wGFvcl06hKZRd+DFAlMheHG466671EuEXoGRI0fSDTfcQNdff716ocA+ufb0CPl/m73cSgk8y+xFu6cipksTK8dPDCAN1QD9xNw2RHuI2qup/WjYJWi/S743EgLPPfccsbvW9rZi360Wr6X3KIDShH26lJ6ebpiSv0ePHuotiN37xFNn9k3Kdx8gwFOQBC8rvAmIJ7F/4dB3KdelHg3/bYPagqdLqHfv3jbPj16b5s2bK28qFoto4ynXnh4h/29rlDLaS4Veo19//VV9BXWJJlaOnxhAGqoB+qlZvvqHp15VrVwLkNbvk+3GQQBTXeAVatWqVYMO4b3RvHO4SWvialxRT8ZWQ6txPj/44ANau3YtXXzxxeqB6qxXV+MnY+cMPev2aYHOXbt2VS8U3377LT3wwAPEsT9q4cEXX3xBMGz14mrsUFfGT4+Yb7cnT56sQjwWLFhA//3vf1XwOoLZ3333XVq8eLG6r+p5E60cP+O1gb49X2ndDQSqqqpUbe2CtD80NjZWFWn17PfLd/8i8MorrxAHSROmVM4880ybMtp4ORpXVJSxtcHl8w0snX799depe/fuavWXqw5djZ+MnSsErdmPVXoQxI3861//IsSTIJYOnh94ThctWkTwLGBFpfaC4mrs0J6MH1BoHMEqvbfffluttHz55ZeJF46ojmG4whMOg1ZPXGvl+IkHqHHG2KNe8APAHDcEQWFGgqXTEOQjEQksBD766CPCH27Gd955p80FDy0xjw1xNK7Yp42t9htAmYj1CODaefDBB9WCAkx9OcoZo/Us16WGhP8/4SmAII8Wpktuu+02FTALQ+jf//43YTk8YvS0hyrqyrUHFAJL1qxZo3I3QSvkb0pKSlIKIj5IHzuJQivHTwwgBXNg/ocHJ/LBQBwZOFp5eHh4YJ5EE9UKN1wk0cMY3n333aSfwwYkeGOFINeFI9HG1j5/iaP6Uu4ZAhgneAuuueYawsohVyLXpSuEGm8/VkpC8FDktARqFRE8PQiKRgA0DFsIpsKQnBQi156CIWD+w8vho48+quIhP/zwQxVPiZV9b731ljJgEdzOaQxs+lo5fmIA2WANzA3NEna0zF0r134UgXkWTUcr3JDhin/vvfdUckrcgPXz1xoS2rgigNORyNg6Qsa6csQY/O9//6MhQ4bQueeea7phbfy0MbI/UCuX69IeGWu/axnW4UlFoLO94MVDiw/SPAna2Mm1Z49W43+Hdw6xP7hOHn74YUpLS7MpgXHD1CWC12HA+mL8JAbIBndgbuBiRYCzdkO111K7iLXcF/b75XvjIYAxwtsKLmrEEODiBa2FkWg3YUfjimNkbI2Qs7ZMy7OEvD7IQaIXzQPHSSoVnQJycL3//vuqilyXeqT8t60ZQO3atXOoBOLv8PBEigOIXHsOoWr0HVjRBRk4cKBtXPRKYFxhxCIXHhYowCiycvzEA6RHOwC3wScF2bFjh6F2WrkzvhTDA6XQUgRgrPz9739Xxg/eYhDM58j4QcfauMK41dz4eoWw/B3L5zHdAt43Ed8gEBYWpmJ/YOzAGNX/aQYQVvKhnDPU2pTQxk+7/mw7ftvQyuW6tEfG2u/aOOA60ic61PcCyhKIxq2nP0auPT1Sjb+t5TjDlKUj0QLStXABK8dPDCBHqAdIucYf9dNPPzXQCMGYnEZclWu5FBpUkgKfI4B8Pv/4xz8UCSpWEcH4QX4fZ4JVDVjhgIfq0qVLG1SdO3euuqGjjrZ6pUElKfAaAfC3YZWQ0R8S6EFwDWK/5i3SyvAp1yVQ8J/06dNHvUwgd5OWm0mvDQxXeH9g6GrGqFx7eoT8u6293CGPHZ5n9oIyEJ5C4P2BWDl+YgApSAP3P7AZZzArLhJ+aflkNG2RMwGEmwjcdJQoUasrn75DABQX69evV65Z5mWz0ZW46lHLNIxgP9yoNcHSXgQDQkCLIRJ4CMh1GRhjAs/BhRdeqJTBcmncDzWBR+j5558nLJvG/VFLhIj9cu1pKPn3s3///pSSkqLGDfdOvEzq5c0331RpRGD06D3qVo2fsMHr0Q7Q7fnz59O9996rPAJMkKmmRPDARQAnbgBg0rVfZRSgpxJyamGa5JxzzrEZMHjTdCR4g0GuGU1wg8bKI8yD4wLH2MIlDK8CbuSjRo1SDPOYBhNpfAQQeInVJ47Y4OW6bPwxMeoR1+Ctt95KWEqNvFq4bhA7ghxAiO3KzMxUhpCWVgJtyLVnhKR/yhAzedNNN6kxwT1y6NChylhdsWKFerFEWopnn3223jPOqvETA8g/Y+52r/iRwF2vBfKhAXiG8MMZNGiQ2+3JAdYgAPcsCE/NCG7Eb3PCL73g5o2HLMgatXgEGFEwqq666qp6b63642Tb9wi4MoCggVyXvh8HMz3ggfjGG2+oFX1arBZWhuHeeOONN9oCZ/VtybWnR8O/2zt37lQcfDBi9YLxu+WWW6hjx476YrVtxfiJAdQA1sAugGcAAX8IBEPCKPEOBPZ4mdUOnp/t27fb+MFk+bRZ5AKjnlyXgTEOmEIBVQKmvRBfYub+KNdeYIwdtEA+J+TkQjZvGD1aHjxnGnozfmIAOUNW9gkCgoAgIAgIAoJASCIgwQUhOaxyUoKAICAICAKCgCDgDAExgJyhI/sEAUFAEBAEBAFBICQREAMoJIdVTkoQEAQEAUFAEBAEnCEgBpAzdGSfICAICAKCgCAgCIQkAmIAheSwykkJAoKAICAICAKCgDMExAByho7sEwQEAUFAEBAEBIGQREAMoJAcVjkpQUAQEAQEAUFAEHCGgBhAztCRfYKAICAICAKCgCAQkgiIARSSwyonJQgIAoKAICAICALOEBADyBk6sk8QEAQEAUFAEBAEQhIBMYBCcljlpAQBQUAQEAQEAUHAGQJiADlDR/YJAoKAIOBjBBYtWuTjHqR5QUAQMEJADCAjVKRMEBAEBAEfI1BYWEgXX3wxTZw40cc9SfOCgCBghIAYQEaoSJkgIAgIAj5G4P7776f333+f6urqfNyTNC8ICAJGCIgBZISKlAkCgoAgIAgIAoJASCMgBlBID6+cnCAgCAgCgoAgIAgYIdDCqFDKBAFBQBDwFwLHjh2jlStXqu579epFrVu3pqNHj9KqVavol19+oYSEBBo/fjxlZmbWU1E7bt68eRQWFkY9e/akyZMnU4sWzm9zmILatm0brV27ljZu3Ejt27enAQMGUN++fSkqKqpeH9oXrS98Rz8xMTFKx3Xr1tHixYupoqKCBg0aRIMHD6akpCTtMPW5d+9e2r17N+3fv199R//Lli1T26mpqdSxY8d69eWLICAI+AgBvvhEBAFBQBAIGAQOHDiAoBj1N3/+/Lrvvvuurk2bNrYybR8bKXVsaCi9FyxYUMeGkWGdzZs3Ozw3Nnzqhg8f3uA49JGcnFz39ddfGx6r13Hu3Ll1v/76a11iYmKDdiIjI+teeeWVem089NBDDepp53TjjTfWqytfBAFBwHcIyBQY33lEBAFBIDAR+PTTT+nss8+m8vJy5ZnBiqm2bdsqZdesWUOXX345zZkzh6ZMmaK8LvDajB07liIiImx1LrvsMsNA4y+++EJ5epYuXUrNmjWjYcOGqfYmTZpEsbGxykMzbdo0uuOOO5yC8+2339Jpp51GJSUl1LVrV8Lxmtfn0KFDdNVVV9Gzzz5ra6Ndu3Y0ZMgQ23mgb3zHX3p6uq2ebAgCgoCPEfCdbSUtCwKCgCDgPgJ67wrf/uri4+PreFrJ1tDBgwfrxo0bZ/OiNG/evK537951PP1kqwPPEE9B2erwtJhtHzZKS0ttHhs2Ours9xcXF9exUWM7Hh4evdjrmJGRUQdPkCbHjx+ve+utt2zHd+7cuY6nzbTd6vPqq69W+9lYq1cuXwQBQaBxEBAPkI8NTGleEBAEPEeAjRv68MMPacSIEbZGEBN077332r5jA8vJ4f3RBDE5L774ovaVtmzZYtvGBo6Hxwbtf/TRR8QGVb39PJ1FPPVG3bt3V+XXX389sVFTr47+Cxs7Ki5JK4NXB56nSy+9VBXt3LmT1q9fr+2WT0FAEAgABMQACoBBEBUEAUHAGAEEEp9++ukNdrLHx1YG4whBy/aiGS8oz83Nte1GAPNrr72mvk+dOpVGjhxp26ffQPA0DB8IArA3bNig323bHjp0aD3jx7aDN/r372/7WlRUZNuWDUFAEPA/As6XR/hfP9FAEBAEmjACRoYN4IiLi7OhYr8aTNsRHR2tbdKRI0ds23l5eXT48GH1HQYKT2fZ9tlvdOvWzVaElWL9+vWzfdc2unTpom02+OSpMVtZbW2tbVs2BAFBwP8IiAHk/zEQDQQBQcABAnpDx0EV4tVahrswDWUkW7dutRU/8sgjhD8zAgPISJwtW9cvo3c2hWbUrpQJAoKAbxGQKTDf4iutCwKCgBcI8DJyL442PhQ5eDyR/Px8w8N8oaNhR1IoCAgCliIgHiBL4ZTGBAFBwEoEHHlxvOkjLS3Ndvi7775L06dPt313thEeHu5st+wTBASBIENADKAgGzBRVxAQBLxDICsry9YAVodhxZiIICAIND0EZAqs6Y25nLEg0KQRQGCyNm2FJIbOYnOQiBGB0kjGyLmALMVN825xxhNL25XGBAFBwBwCYgCZw0lqCQKCQIgggNw/N9xwgzobZJN+6aWXDM+ssvOp/C4AAAK7SURBVLJS5QsCv9fs2bOpR48ehvU8LdSMMKxQQ18igoAg0LgIiAHUuHhLb4KAIBAACNx9993UoUMHpQmMobvuuouqqqrUdyyRnzVrFp1xxhmUnZ2tyq699lobvYVV6nOGa1tTDzzwADGfGW3atMlWJhuCgCDgWwTEAPItvtK6ICAIBCACyCaNDNDg3kJiRCyFB/8XvDzg8YLxA4MEMmPGDHr88cctPwtkn9amwZ588knFYfbPf/7T8n6kQUFAEDBGQAwgY1ykVBAQBEIcgdGjR6vszldeeaUyfhALhKBobTqK+btUxuiPP/6YwsLCLEcDpK3PP/+8InnVGmfmem1TPgUBQcDHCDQD5ZiP+5DmBQFBQBAIeASQ52fjxo0qQBrZnbFcHvFCjSEFBQVUXV2t+tSY7BujX+lDEGjKCIgB1JRHX85dEBAEBAFBQBBoogg0zutNEwVXTlsQEAQEAUFAEBAEAhMBMYACc1xEK0FAEBAEBAFBQBDwIQJiAPkQXGlaEBAEBAFBQBAQBAITATGAAnNcRCtBQBAQBAQBQUAQ8CECYgD5EFxpWhAQBAQBQUAQEAQCEwExgAJzXEQrQUAQEAQEAUFAEPAhAmIA+RBcaVoQEAQEAUFAEBAEAhMBMYACc1xEK0FAEBAEBAFBQBDwIQJiAPkQXGlaEBAEBAFBQBAQBAITATGAAnNcRCtBQBAQBAQBQUAQ8CECYgD5EFxpWhAQBAQBQUAQEAQCEwExgAJzXEQrQUAQEAQEAUFAEPAhAmIA+RBcaVoQEAQEAUFAEBAEAhMBMYACc1xEK0FAEBAEBAFBQBDwIQJiAPkQXGlaEBAEBAFBQBAQBAITATGAAnNcRCtBQBAQBAQBQUAQ8CECYgD5EFxpWhAQBAQBQUAQEAQCEwExgAJzXEQrQUAQEAQEAUFAEPAhAv8PzxZrpGJTLO4AAAAASUVORK5CYII=)
推定結果の評価
logLik(npub.poisson1)
'log Lik.' -1651.056 (df=6)
- 自由度909での5%水準の\(\chi^2\)値は、以下のように得られます。
qchisq(0.95, df.residual(npub.poisson1))
[1] 980.2518
- モデルがデータに当てはまるかどうかをみるために、devianceとPearson’s \(\chi^2\)値を計算します。
- Deviance
npub.poisson1$deviance
[1] 1634.371
sum((residuals(npub.poisson1,"pearson"))^2)
[1] 1662.547
- これらの結果から、モデルは当てはまりが良くないことがわかります。
- 過分散とdispersion parameterを以下のように計算します。
- 過分散
npub.poisson1$deviance/npub.poisson1$df.res
[1] 1.797988
sum(residuals(npub.poisson1, type="pearson")^2)/npub.poisson1$df.res
[1] 1.828984
モデル選択
- まずANOVAを用いて、モデル
npub.poisson1
の推定結果を評価します。
- ANOVA(Type I)は
anova()
関数を用いて計算します。
anova(npub.poisson1, test="Chisq")
Analysis of Deviance Table
Model: poisson, link: log
Response: art
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 914 1817.4
fem 1 23.029 913 1794.4 1.596e-06 ***
mar 1 0.251 912 1794.1 0.616719
kid5 1 17.388 911 1776.7 3.047e-05 ***
phd 1 10.499 910 1766.2 0.001194 **
ment 1 131.868 909 1634.4 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
- Type IのANOVAでは、変数
mar
に効果がないと、誤って検出されてしまいます。
- Type IIとType IIIのANOVAは
car
パッケージのAnova()
関数を用いて研鑽します。
- ANOVA(TypeII)
car::Anova(npub.poisson1, type=c("II"))
Analysis of Deviance Table (Type II tests)
Response: art
LR Chisq Df Pr(>Chisq)
fem 17.075 1 3.593e-05 ***
mar 6.431 1 0.01121 *
kid5 22.082 1 2.613e-06 ***
phd 0.236 1 0.62696
ment 131.868 1 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
car::Anova(npub.poisson1, type=c("III"))
Analysis of Deviance Table (Type III tests)
Response: art
LR Chisq Df Pr(>Chisq)
fem 17.075 1 3.593e-05 ***
mar 6.431 1 0.01121 *
kid5 22.082 1 2.613e-06 ***
phd 0.236 1 0.62696
ment 131.868 1 < 2.2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
- 次に、変数
phd
を減らしたモデルnpub.poisson2
を推定します。
npub.poisson2 <- glm(art~fem+mar+kid5+ment, data=npub,
family = poisson(link = "log"))
summary(npub.poisson2)
Call:
glm(formula = art ~ fem + mar + kid5 + ment, family = poisson(link = "log"),
data = npub)
Deviance Residuals:
Min 1Q Median 3Q Max
-3.6436 -1.5408 -0.3583 0.5623 5.3986
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.34517 0.06012 5.741 9.41e-09 ***
femWomen -0.22530 0.05461 -4.125 3.70e-05 ***
marMarried 0.15218 0.06107 2.492 0.0127 *
kid5 -0.18499 0.04014 -4.609 4.05e-06 ***
ment 0.02576 0.00195 13.212 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for poisson family taken to be 1)
Null deviance: 1817.4 on 914 degrees of freedom
Residual deviance: 1634.6 on 910 degrees of freedom
AIC: 3312.3
Number of Fisher Scoring iterations: 5
npub.poisson1
とnpub.poisson2
のAICを比較します。
c(AIC.model1=AIC(npub.poisson1), AIC.model2=AIC(npub.poisson2))
AIC.model1 AIC.model2
3314.113 3312.349
- ANOVAにより
npub.poisson1
とnpub.poisson2
のAICを比較します。
anova(npub.poisson1, npub.poisson2, test="Chisq")
Analysis of Deviance Table
Model 1: art ~ fem + mar + kid5 + phd + ment
Model 2: art ~ fem + mar + kid5 + ment
Resid. Df Resid. Dev Df Deviance Pr(>Chi)
1 909 1634.4
2 910 1634.6 -1 -0.2362 0.627
参考:Robust SE
result <- npub.poisson2
cov.result <- vcovHC(result, type="HC0")
std.err <- sqrt(diag(cov.result))
r.est <- cbind(Estimate= coef(result), "Robust SE" = std.err,
"Pr(>|z|)" = 2 * pnorm(abs(coef(result)/std.err), lower.tail=FALSE),
LL = coef(result) - 1.96 * std.err,
UL = coef(result) + 1.96 * std.err)
r.est
Estimate Robust SE Pr(>|z|) LL UL
(Intercept) 0.34517404 0.082177620 2.665130e-05 0.18410591 0.50624218
femWomen -0.22530260 0.071201599 1.554611e-03 -0.36485773 -0.08574746
marMarried 0.15217527 0.083642215 6.885659e-02 -0.01176348 0.31611401
kid5 -0.18499262 0.055799227 9.153637e-04 -0.29435910 -0.07562613
ment 0.02576117 0.003491732 1.609692e-13 0.01891738 0.03260497