離散選擇模型-資料下載頁(yè)
2025-09-25 19:01本頁(yè)面
【正文】 est exp2 ? ( 1) exp2=0。chi2(1)=。Prob chi2 = ? . test potexp=grade ? ( 1) potexp grade = 0 ? chi2( 1) = 。Prob chi2 = Logit model ? logit depvar [indepvars][weight][if exp][in range][,nocoef noconstant robust cluster(varname) score(newvarlist) level() maximize_options]. ? predict [type] newvarname [if exp][in range][,statistic rules asif nooffset] ? Statistic is : ? {p|xb|stdp}as the same as probit. Predict ? dbeta :Pregibon(1991) influence statistic. ? deviance:deviation residual。越大，模型越差； ? dx2:Hosmer,and Leme influence statistic. ? ddeviance:Hosmer,Leme influence statistic. ? hat:pregibon,leverage. ? residuals,pearson residuals， ? rstandard:standard pearson residuals?；驑?biāo)準(zhǔn)化殘差 。大樣本時(shí)服從正態(tài)分布 N(0,1). ? or:reports estimated coefficients transformed into odds (beta) instead of beta。 ??2??D?Mlogit procedure ? mlogit depvar [indepvars][weight][if exp][in range][,basecategory() constraints(clist) noconstant robust cluster(varname) score(newvarlist) rrr level() maximize_options]. ? predict [type] newvarname [if exp][in range][,{p|xb|stdp|stddp }oute(oute). explanation ? basecategory():specifies the value of depvar that treated as base category。default is to choose the most frequent category. ? constraints(clist):linear constrained estimation. ? score(newvarlist):create k1 variables,k is number of observed outes。the mth is dLnLj/dlnxjbm。 ? Predict:p,xb,stdp as before。stdpp:calculates standard error of the difference in two linear predictions,must specifies oute(oute)。 ? :predict sed1,stdpp(1,3) ? rrr:reports estimated coefficients transformed to relative risk :exp(b) instead of b。 例子：加入不同行業(yè) ? 數(shù)據(jù)仍然使用前面工會(huì)例子所使用的 。若選區(qū)第一類產(chǎn)業(yè)（自然資源）作為基準(zhǔn)； ? 因變量：產(chǎn)業(yè)選擇 ind1。 ? 自變量：potexp,exp2,grade,married,union,high ? Multinomial logit model results: 輸出結(jié)果 ? ……………………………………………………… ..省略 ? 13 | ? potexp | .0665262 ? exp2 | .004946 .0015428 .0019222 .0079698 ? union | .7321836 .8778095 ? grade | .376377 .0877074 .2044736 .5482805 ? high | 4738540 9287427 9287308 ? married | .4845398 .1065695 ? _cons | ? + ? 14 | ? potexp | .1043477 .102058 ? exp2 | .0023012 .0021146 .0064458 ? union | .47805 .5203087 ? grade | .3177647 .1038616 .1141998 .5213296 ? high | .5012028 ? married | .7061995 .3233226 ? _cons | . . . . . ? (Oute ind1==1 is the parison group) 例子：選擇行業(yè) (logit… .,rrr） ? …… rrr (odds ratio) ? + ? 14 | ? potexp | .8988697 .0933506 .7333253 ? exp2 | .0021047 .9978678 ? union | .6549112 .4726132 ? high | .1217621 ? married | .4224411 .295379 .1072975 ? ? (Oute ind1==1 is the parison group) 其他過(guò)程 ? Ordered logit: ologit ? Conditional logistic regression:clogit ? Weighted least squares for grouped data:glogit。gprobit。 ? Nested logit model:nlogit ? Panel logit/probit:xtlogit,xtprobit ? Sas/stat/ proc logistic。 ? 推薦參考書(shū)： ? 1。 Logistic 回歸模型方法與應(yīng)用 。高教出版社。 ? 2。 Statistics with ? of stata。
語(yǔ)文相關(guān)推薦
鄂ICP備17016276號(hào)-1