Analyzing Social Networks Experiments in R The 4th China-R Conference, ECNU, Shanghai
Correlation v.s. Casuality Nobel Prize: VAR
The object of science is the discovery of relations.., of which the complex may be deduced from the simple. John Pringle Nichol, 1840 Data Mining, Econometrics, Other Statistical Models. Handbook of Econometrics
The object of science is the discovery of relations.., of which the complex may be deduced from the simple. John Pringle Nichol, 1840 Data Mining, Econometrics, Other Statistical Models. Handbook of Econometrics
The object of science is the discovery of relations.., of which the complex may be deduced from the simple. John Pringle Nichol, 1840 Data Mining, Econometrics, Other Statistical Models. Handbook of Econometrics
self-selection Network Formation Models peer effects Social Learning ModelsInformation Spread Models, Social Influencer Models
Source: Wang, Alex, Xiaoquan (Michael) Zhang and Il-Horn Hann, 2010, Social Bias in Online Product Ratings Workshop on Information Systems and Economics (WISE), December 2010, St. Louis, USA.
Fisher
0121 345678 98678
103180 179635060 38674627009 151 78 1084111 2 173 96 122150 155157 116109 58 184187 377127 576690 9302 1746123185012023 190101 1821412224 1815156 188117 168160 135121172 162165 102194 131139130 163 177164106159 147143 105193 11477404472 176161 197134154 142144332825 0007 196148169 6 1215 54484245 153128127 146 126183 0536 170 104120 1494347 7399595352 186118191 3935 119110111115 199 113192301917136 175169865 7 684675133 6421107138195 2679 149145959794 156140101671251893 112
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 10 20 30 40 50 60 t = 50 Repetitions = 20 N = 198
DIDDifference in Difference) with Probit/Logit Models Pre-introduction Post-introduction YA-YB=-1 more suitable: A introduce potatoes A Population Population YA-YB=1 B not suitable no potato introduction B Population Population DD=2, caused by potatoes
Infected = α + β 1 Network + β 2 Traditional + β 3 NT + ε data.table lm() glm(); VGAM package - probit() logit() mylogit<- glm(infected~network+traditional+nt, family=binomial(link="logit"), na.action=na.pass, data=mydata) summary(mylogit) library(aod) wald.test.. #VGAM for multinominal cases fit.ms <- vgam(infected_level ~ network+traditional+nt, multinomial(reflevel=1), data = nzmarital)