![\"\"](\"https:\/\/info.zanet.biz\/wp\/wp-content\/uploads\/2023\/05\/virtualdat.png\")
\u3053\u306e\u3088\u3046\u306a\u30c7\u30fc\u30bf\u3092Excel\u306a\u3069\u3067\u7528\u610f\u3057\u3066\u304a\u304d\u307e\u3059\u3002\u306a\u304a\u3001\u3053\u3053\u306b\u793a\u3059\u306e\u306f\u67b6\u7a7a\u306e\u30c7\u30fc\u30bf\u3067\u3059\u3002<\/p>\n\n\n\n
nnet\u306fBrian Ripley, William Venables\u306b\u3088\u308bR\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u3067\u3059\u3002 Link<\/a> <\/p>\n\n\n\n <\/p>\n\n\n\n <\/p>\n\n\n\n <\/p>\n\n\n\n \u4ee5\u4e0bR\u3067\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u5b9f\u884c\u3057\u3066\u3044\u304d\u307e\u3059\u3002\u5909\u6570\u540d\u306f\u56f31\u306e\u3082\u306e\u3092\u7528\u3044\u3066\u3044\u307e\u3059\u304c\u3001\u5b9f\u969b\u306b\u56f31\u306e\u30c7\u30fc\u30bf\u3092\u89e3\u6790\u3057\u3066\u3044\u308b\u308f\u3051\u3067\u306f\u3042\u308a\u307e\u305b\u3093\u3002\u30b9\u30af\u30ea\u30d7\u30c8\u3068\u624b\u9806\u3092\u793a\u3059\u306e\u304c\u76ee\u7684\u3067\u3059\u3002\u305d\u308c\u305e\u308c\u81ea\u5206\u306e\u30c7\u30fc\u30bf\u3067\u5fc5\u8981\u306b\u5fdc\u3058\u3066\u5909\u6570\u540d\u306a\u3069\u5909\u66f4\u3057\u3066\u5b9f\u884c\u3057\u3066\u304f\u3060\u3055\u3044\u3002<\/p>\n\n\n\n packneed=c(“nnet”,”tcltk2″);current=installed.packages();addpack=setdiff(packneed,rownames(current));url=”https:\/\/cran.ism.ac.jp\/”;if(length(addpack)>0){install.packages(addpack,repos=url)};if(length(addpack)==0){print(“Already installed.”)} library(nnet) train.dat=read.delim(“clipboard”,sep=”\\t”,header=TRUE) train.dat$Diagnosis = as.factor(train.dat$Diagnosis) multinom_model = multinom(Diagnosis ~ Test1+Test2+Test3, data = train.dat) summary(multinom_model)<\/p>\n\n\n\n z=summary(multinom_model)$coefficients\/summary(multinom_model)$standard.errors exp(coef(multinom_model))<\/p>\n\n\n\n (predDiagnosis=predict(multinom_model, newdata=train.dat, type=”probs”)) results=cbind(train.dat,predDiagnosis)<\/p>\n\n\n\n filnam=tclvalue(tkgetSaveFile(initialfile=”.txt”,filetypes=”{{Text Files} {.txt}} {{All files} *}”));if(filnam!=””){write.table(results,filnam,sep=”\\t”,row.names=FALSE,col.names=TRUE)} Test1=3 (CasePredicted=predict(multinom_model, newdata = case, type=”class”)) (CasePredicted=predict(multinom_model, newdata = case, type=”probs”)) \u8a3a\u65ad\u306ePrediction\u306f\u4e0a\u8a18\u306epredict()\u95a2\u6570\u3092\u7528\u3044\u3066\u3067\u304d\u307e\u3059\u304c\u3001\u81ea\u5206\u3067\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u66f8\u3044\u3066\u8a08\u7b97\u3057\u305f\u3044\u5834\u5408\u306f\u4ee5\u4e0b\u306e\u69d8\u306b\u3057\u307e\u3059\u3002 \u691c\u67fb\u304c3\u3064\u3001\u3059\u306a\u308f\u3061\u3001\u8aac\u660e\u5909\u6570\u304c3\u3064\u306e\u5834\u5408\u3067\u3001\u305d\u306e\u75c7\u4f8b\u306e\u691c\u67fb\u7d50\u679c\u304cx1, x2, x3\u3067\u3042\u3042\u308c\u3070\u3001 \u901a\u5e38\u306eLogistic regression\u306e\u5834\u5408\u306f\u3001y = a + b1*x1 + b2*x2 + b3*x3\u3067\u3001 \u6587\u732e<\/strong>\uff1a \u5fdc\u7b54\u5909\u6570\uff08\u5f93\u5c5e\u5909\u6570\uff09\u304c\u4e8c\u5024\u5909\u6570\u3001\u4f8b\u3048\u3070\u3042\u308b\u75be\u60a3\u30fb\u305d\u3046\u3067\u306a\u3044\u3067\u3001\u3044\u304f\u3064\u304b\u306e\u8aac\u660e\u5909\u6570\u304b\u3089\u305d\u308c\u3092\u63a8\u5b9a\u3059\u308b\u30e2\u30c7\u30eb\u3092\u4f5c\u6210\u3057\u305f\u3044\u5834\u5408\u3001\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u3092\u7528\u3044\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u8aac\u660e\u5909\u6570\u306f\u9023\u7d9a\u5909\u6570\u3067\u3082\u540d\u7fa9\u5909\u6570\u3067\u3082\u4f7f\u3048\u307e\u3059\u3002 \u5fdc\u7b54\u5909 … “Multi-nomial logistic regression” \u306e<\/span>\u7d9a\u304d\u3092\u8aad\u3080<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[32],"tags":[],"class_list":["post-1380","post","type-post","status-publish","format-standard","hentry","category-statistics"],"_links":{"self":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/1380","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1380"}],"version-history":[{"count":4,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/1380\/revisions"}],"predecessor-version":[{"id":1386,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/1380\/revisions\/1386"}],"wp:attachment":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1380"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1380"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1380"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}#Install packages needed.<\/h6>\n\n\n\n
\u4eca\u56de\u53d6\u308a\u4e0a\u3052\u305f\u30d1\u30c3\u30b1\u30fc\u30b8\u4ee5\u5916\u306b\u3001NeuralNetTools\u3001EffectStars2\u3068\u3044\u3046\u30d1\u30c3\u30b1\u30fc\u30b8\u3082\u6709\u7528\u3067\u3059\u3002\u4eca\u56deMulti-nomial logistic regression analysis\u3092nnet\u30d1\u30c3\u30b1\u30fc\u30b8\u306emultinom()\u95a2\u6570\u3067\u884c\u3063\u3066\u3044\u307e\u3059\u304c\u3001nnet\u306f\u305d\u308c\u3060\u3051\u3067\u306a\u304f\u3001Neural network\u3092\u7528\u3044\u308b\u89e3\u6790\u3082\u3067\u304d\u307e\u3059\u3002<\/p>\n\n\n\n#Read in packages.<\/h6>\n\n\n\n
library(tcltk2)<\/p>\n\n\n\n#Read in data.<\/h6>\n\n\n\n
#Excel\u3067\u5fc5\u8981\u306a\u30c7\u30fc\u30bf\u7bc4\u56f2\u3092\u30b3\u30d4\u30fc\u3057\u3066\u304b\u3089R\u306b\u623b\u308a\u3053\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u5b9f\u884c\u3057\u307e\u3059\u3002\u30af\u30ea\u30c3\u30d7\u30dc\u30fc\u30c9\u3092\u4ecb\u3057\u3066\u3001\u30c7\u30fc\u30bf\u304c\u5909\u6570train.dat\u306b\u683c\u7d0d\u3055\u308c\u307e\u3059\u3002Mac\u306e\u5834\u5408\u306f\u3001train.dat=read.delim(pipe(“pbpaste”),sep=”\\t”,header=TRUE) \u3067\u3059\u3002<\/p>\n\n\n\n#Set a reference as a factor variable.<\/h6>\n\n\n\n
train.dat$Diagnosis = relevel(train.dat$Diagnosis, ref = “D1”)
\u30d9\u30fc\u30b9\u3068\u306a\u308bDiagnosis\u3092\u8a2d\u5b9a\u3057\u307e\u3059\u3002\u305d\u306e\u969b\u306b\u3001\u6570\u5024\u3067\u306f\u306a\u304ffactor\u3068\u3057\u3066\u53d6\u308a\u6271\u3046\u306e\u3067\u3001as.factor()\u95a2\u6570\u3092\u7528\u3044\u3066\u3044\u307e\u3059\u3002ref\u5f15\u6570\u3067\u8a2d\u5b9a\u3059\u308bDiagnosis\u306f\u901a\u5e38\u306f\u5bfe\u7167\u3068\u306a\u308b\u3088\u3046\u306a\u8a3a\u65ad\u30fb\u75c5\u614b\u3067\u3059\u3002<\/p>\n\n\n\n#Multi-nomial logistic regression analysis.<\/h6>\n\n\n\n
\u8a3a\u65ad ~ \u8aac\u660e\u5909\u65701 + \u8aac\u660e\u5909\u65702 + \u8aac\u660e\u5909\u65703\u306e\u69d8\u306bformula\u3092\u8a18\u8ff0\u3057\u307e\u3059\u3002<\/p>\n\n\n\n#Show the results: Intercepts, coefficients for explanatory variables.<\/h6>\n\n\n\n
#P values for explanatory variables.<\/h6>\n\n\n\n
P_value=(1-pnorm(abs(z),0,1))*2
P_value
summary()\u95a2\u6570\u3067\u306fP\u5024\u306f\u8a08\u7b97\u3055\u308c\u306a\u3044\u306e\u3067\u3001\u3053\u306e\u3088\u3046\u306a\u30b9\u30af\u30ea\u30d7\u30c8\u3067\u8a08\u7b97\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n\n\n\n#Odds ratios for explanatory variables (exponentials of the coefficients).<\/h6>\n\n\n\n
#Prediction the diagnosis with explanatory variables for each case with the training data.<\/h6>\n\n\n\n
\u3053\u3053\u3067\u306f\u3001\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u306b\u7528\u3044\u305f\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u3001Prediction\u3092\u884c\u3063\u3066\u3044\u307e\u3059\u3002\u9069\u5408\u5ea6\u304c\u9ad8\u3044\u306e\u3067\u3001\u6b63\u8a3a\u7387\u306f\u9ad8\u304f\u306a\u308a\u307e\u3059\u3002 <\/p>\n\n\n\n#Prepare the result data to be saved.<\/h6>\n\n\n\n
#Save the result data as a text file (tab-separated).<\/h6>\n\n\n\n
\u30d5\u30a1\u30a4\u30eb\u540d\u3092\u8a2d\u5b9a\u3057\u3066\u4efb\u610f\u306e\u30d5\u30a9\u30eb\u30c0\u30fc\u306b\u30c6\u30ad\u30b9\u30c8\u30d5\u30a1\u30a4\u30eb\u3068\u3057\u3066\u3001\u5143\u306e\u30c7\u30fc\u30bf\u3068Prediction\u306e\u7d50\u679c\u3092\u5408\u308f\u305b\u3066\u30bf\u30d6\u533a\u5207\u308a\u306e\u30c6\u30fc\u30d6\u30eb\u306e\u5f62\u5f0f\u3067\u4fdd\u5b58\u3057\u307e\u3059\u3002<\/p>\n\n\n\n#Calcuating disease probabilities for individual explanatory variable data.
Set data for the case:
an example<\/h6>\n\n\n\n
Test2=14
Test3=5
case=cbind(Test1,Test2,Test3)
row.names(case)[1]=”case”
\u4efb\u610f\u306e\u691c\u67fb\u7d50\u679c\u306e\u5834\u5408\u306e\u3001\u8a3a\u65ad\u306e\u63a8\u5b9a\u3092\u884c\u3044\u307e\u3059\u3002<\/p>\n\n\n\n#Output Diagnosis class.<\/h6>\n\n\n\n
3\u3064\u306e\u8a3a\u65ad\u306e\u5185\u3001\u75be\u60a3\u78ba\u7387\u304c\u4e00\u756a\u9ad8\u3044\u3082\u306e\u304c\u51fa\u529b\u3055\u308c\u307e\u3059\u3002<\/p>\n\n\n\n#Output probabilities.<\/h6>\n\n\n\n
3\u3064\u306e\u8a3a\u65ad\u306e\u305d\u308c\u305e\u308c\u306e\u75be\u60a3\u78ba\u7387\u304c\u51fa\u529b\u3055\u308c\u307e\u3059\u3002<\/p>\n\n\n\n
Multi-nomial logistic regression\u306e\u7d50\u679c\u3001Reference diagnosis\u4ee5\u5916\u306eDiagnosis\u306b\u5bfe\u3059\u308b\u5207\u7247\u3001\u4fc2\u6570\u306e\u5024\u304c\u5f97\u3089\u308c\u307e\u3059\u3002\u3053\u308c\u3089\u306e\u5024\u3092\u7528\u3044\u3066\u3001\u5404\u75c7\u4f8b\u306e\u75be\u60a3\u78ba\u7387\u3092\u8a08\u7b97\u3057\u307e\u3059\u3002\u5207\u7247\u3001\u4fc2\u6570\u306e\u5024\u306f\u3001coef(multinom_model)\u3067\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u3068\u3057\u3066\u5f97\u3089\u308c\u307e\u3059\u3002\u305d\u3053\u304b\u3089\u3001\u5fc5\u8981\u306a\u5207\u7247\u3001\u4fc2\u6570\u306e\u5024\u3092\u8aad\u307f\u51fa\u3057\u3066\u3001\u305d\u306e\u75c7\u4f8b\u306e\u305d\u308c\u305e\u308c\u306e\u691c\u67fb\u7d50\u679c\u306e\u5024\u306b\u305d\u308c\u305e\u308c\u306e\u5bfe\u5fdc\u3059\u308b\u4fc2\u6570\u3092\u639b\u3051\u7b97\u3057\u3066\u3001\u7dcf\u548c\u3092\u6c42\u3081\u3001\u3055\u3089\u306b\u5207\u7247\u306e\u5024\u3092\u52a0\u7b97\u3057\u307e\u3059\u3002\u5f97\u3089\u308c\u305f\u5024\u306eExponential\u3092\u6c42\u3081\u307e\u3059\u3002\u305d\u308c\u3092\u5404Diagnosis\u306b\u3064\u3044\u3066\u8a08\u7b97\u3057\u3001\u305d\u308c\u3089\u306e\u7dcf\u548c\u3092\u6c42\u3081\u3001\u3055\u3089\u306b1\u3092\u52a0\u7b97\u3057\u307e\u3059\u3002\u305d\u306e\u5024\u3067\u3001\u5404Diagnosis\u306eExponential\u306e\u5024\u3092\u5272\u308a\u7b97\u3059\u308b\u3068\u3001\u5404Diagnosis\u306b\u5bfe\u3059\u308b\u75be\u60a3\u78ba\u7387\u304c\u5f97\u3089\u308c\u307e\u3059\u3002\u52a0\u7b97\u3059\u308b1\u306fReference diagnosis\u306e\u5207\u7247\u3068\u4fc2\u6570\u30920\u306b\u8a2d\u5b9a\u3059\u308b\u305f\u3081\u3067\u30010\u306eExponential\u304c1\u306b\u306a\u308b\u305f\u3081\u3067\u3059\u3002<\/p>\n\n\n\n
\u8a3a\u65adi\u306b\u3064\u3044\u3066\u3001yi = ai + b1i*x1 + b2i*x2 + b3i*x3
P(\u8a3a\u65adi) = exp(yi)\/(1 + \u03a3exp(yi))<\/p>\n\n\n\n
P(\u8a3a\u65ad)=1\/(1 + exp(-y)) \u3067\u3059\u304b\u3089\u8a08\u7b97\u6cd5\u304c\u9055\u3044\u307e\u3059\u3002<\/p>\n\n\n\n
Ciaburro G, Venkateswaran B: Neural Networks with R<\/strong>: Smart models using CNN, RNN, deep learning, and artificial intelligence principles. 2017, Packt Publishing, Birmingham, UK.
Xu J: Modern Applied Regression<\/strong>. 2023, CRC press, Taylor & Francis Group.
Multi-nomial logistic regression with R<\/strong>. R-bloggers.<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"