{"id":1642,"date":"2025-05-23T21:29:11","date_gmt":"2025-05-23T12:29:11","guid":{"rendered":"https:\/\/info.zanet.biz\/?p=1642"},"modified":"2025-05-24T18:10:11","modified_gmt":"2025-05-24T09:10:11","slug":"%e5%a4%9a%e9%a0%85%e3%83%ad%e3%82%b8%e3%82%b9%e3%83%86%e3%82%a3%e3%83%83%e3%82%af%e5%9b%9e%e5%b8%b0%e5%88%86%e6%9e%90-2-google-gemni%e3%81%ab%e8%81%9e%e3%81%8d%e3%81%aa%e3%81%8c%e3%82%89nnet","status":"publish","type":"post","link":"https:\/\/info.zanet.biz\/?p=1642","title":{"rendered":"\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790 – 2 – Google Gemini\u306b\u805e\u304d\u306a\u304c\u3089nnet\u3067\u89e3\u6790"},"content":{"rendered":"\n

Google Gemini\u3092\u4f7f\u3063\u3066\u3001\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u3092\u3055\u3089\u306b\u63a2\u7d22\u3057\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/p>\n\n\n\n

R\u306e\u30d1\u30c3\u30b1\u30fc\u30b8nnet<\/a>\u3092\u7528\u3044\u3066\u3001\u533b\u7642\u306e\u30c7\u30fc\u30bf\u3067\u306f\u3042\u308a\u307e\u305b\u3093\u304c\u30b5\u30f3\u30d7\u30eb\u30c7\u30fc\u30bf\u3067\u5b9f\u969b\u306b\u5206\u6790\u3092\u5b9f\u884c\u3057\u3001\u8aac\u660e\u5909\u6570\u306e\u5024\u304b\u3089\u63a8\u5b9a\u3059\u308b\u3068\u3053\u308d\u307e\u3067\u3001Gemini 2.5 Flash\u306b\u805e\u304d\u306a\u304c\u3089\u3001\u3084\u3063\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

\u307e\u305a\u3001\u8a66\u3057\u306b\u4f7f\u3048\u308b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u304c\u7121\u3044\u304b\u805e\u3044\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

R\u30d1\u30c3\u30b1\u30fc\u30b8\u306ennet\u306b\u4ed8\u5c5e\u3057\u3066\u3044\u308b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306f\uff1f<\/mark><\/p>\n\n\n\n

\u4ed8\u5c5e\u3059\u308b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306f\u306a\u3044\u3068\u3044\u3046\u56de\u7b54\u3067\u3057\u305f\u3002 \u305d\u306e\u4ee3\u308f\u308a\u306b R \u306b\u6a19\u6e96\u3067\u4ed8\u5c5e\u3057\u3066\u3044\u308b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u304c\u3044\u304f\u3064\u304b \u30ea\u30b9\u30c8\u30a2\u30c3\u30d7\u3055\u308c\u307e\u3057\u305f\u3002 iris\u306f\u3001\u30a2\u30e4\u30e1\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u3001\u5206\u985e\u554f\u984c\u306e\u4f8b\u3068\u3057\u3066\u3088\u304f\u4f7f\u308f\u308c\u307e\u3059\u3068\u306e\u3053\u3068\u3067\u3059\u3002<\/p>\n\n\n\n

iris\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u5909\u6570dat\u306b\u683c\u7d0d\u3059\u308bR\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u306f\uff1f<\/mark><\/p>\n\n\n\n

dat <- iris<\/p>\n\n\n\n

\u305d\u3053\u3067\u6b21\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u5b9f\u884c\u3057\u3001\u3069\u306e\u3088\u3046\u306a\u30c7\u30fc\u30bf\u304b\u78ba\u8a8d\u3057\u307e\u3057\u305f\u3002\u30b3\u30f3\u30bd\u30fc\u30eb\u3078\u306e\u51fa\u529b\u3067\u3059\u3002<-\u3068=\u306f\u540c\u3058\u3067\u3001\u53f3\u5074\u306e\u5909\u6570\u306e\u30c7\u30fc\u30bf\u3092\u5de6\u5074\u306e\u5909\u6570\u306b\u683c\u7d0d\u3057\u307e\u3059\u3002\u30c7\u30fc\u30bf\u578b\u3082\u540c\u3058\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u3053\u3053\u3067\u306f\u3001\u30b5\u30f3\u30d7\u30eb1\uff5e4\u307e\u3067\u306e\u30c7\u30fc\u30bf\u3092\u793a\u3057\u307e\u3059\u3002\u5b9f\u969b\u306b\u306f\u30a2\u30e4\u30e1150\u500b\u4f53\u306e\u30c7\u30fc\u30bf\u3067\u3059\u3002\u9023\u7d9a\u5909\u6570\u306e\u8aac\u660e\u5909\u6570\u304c4\u7a2e\u985e\u3001\u5f93\u5c5e\u5909\u6570\u306fSpecies\u3067\u3001\u30a2\u30e4\u30e1\u306e\u7a2e\u985e\u306e\u69d8\u3067\u3059\u3002\u30a2\u30e4\u30e1\u306e\u7a2e\u985e\u306f\u3044\u304f\u3064\u3042\u308b\u306e\u304b\u8abf\u3079\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/p>\n\n\n\n

\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8dat\u306eSpecies\u306e\u5c5e\u6027\u306f\u3044\u304f\u3064\u3042\u308b\u304b\u8abf\u3079\u308b\u306b\u306f\uff1f<\/mark><\/p>\n\n\n\n

Species\u306f\u6587\u5b57\u5217\u306e\u540d\u7fa9\u5909\u6570\u3067\u3001nlevels(dat$Species)\u3067\u4f55\u7a2e\u985e\u542b\u307e\u308c\u3066\u3044\u308b\u304b\u3092\u77e5\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3088\u3046\u3067\u3059\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\uff13\u7a2e\u985e\u3067\u3042\u308b\u3053\u3068\u304c\u308f\u304b\u308a\u307e\u3057\u305f\u3002<\/p>\n\n\n\n

\u540d\u7fa9\u5909\u6570\u306e\u305d\u308c\u305e\u308c\u306e\u5024\u306flevels\u3068\u8868\u73fe\u3059\u308b\u3088\u3046\u306a\u306e\u3067\u3001\u305d\u308c\u305e\u308c\u306e\u540d\u7fa9\u5909\u6570\u306e\u5024\u3092\u77e5\u308b\u306b\u306f\u3069\u3046\u3059\u308b\u304b\u3092\u805e\u3044\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

Species\u306e\u30ec\u30d9\u30eb\u3092\u77e5\u308b\u306b\u306f\uff1f<\/mark><\/p>\n\n\n\n

levels(dat$Species)\u3067\u308f\u304b\u308b\u3088\u3046\u3067\u3059\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

setosa, versicolor, virginica\u306e3\u7a2e\u985e\u3067\u3042\u308b\u3053\u3068\u304c\u78ba\u8a8d\u3067\u304d\u307e\u3057\u305f\u3002\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0dat\u306e\u30ab\u30e9\u30e0\u540d\u3092\u51fa\u529b\u3055\u305b\u305f\u3044\u3068\u601d\u3044\u307e\u3059\u3002<\/p>\n\n\n\n

dat\u306e\u30ab\u30e9\u30e0\u540d\u3092\u51fa\u529b\u3059\u308b\u306b\u306f\uff1f<\/mark><\/p>\n\n\n\n

colnames(dat)\u3067\u3001\u3042\u308b\u3044\u306fnames(dat)\u3067\u30ab\u30e9\u30e0\u540d\u3092\u51fa\u529b\u3067\u304d\u308b\u305d\u3046\u3067\u3059\u3002<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u3055\u3066\u3001\u4ee5\u4e0a\u304b\u3089\u3053\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8iris\u306f”Sepal.Length”\u3001”Sepal.Width”\u3001 “Petal.Length” “Petal.Width”\u306e4\u3064\u306e\u8aac\u660e\u5909\u6570\u304b\u3089”Species”\u3092\u63a8\u5b9a\u3059\u308b\u305f\u3081\u306b\u3001\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u3092\u9069\u7528\u3067\u304d\u305d\u3046\u306a\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3057\u305f\u3002nnet\u3092\u4f7f\u3063\u3066\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u3092\u5b9f\u884c\u3057\u3066\u307f\u3088\u3046\u3068\u601d\u3044\u307e\u3059\u3002<\/p>\n\n\n\n

R\u306e\u30d1\u30b1\u30fc\u30b8nnet\u3092\u4f7f\u3063\u3066\u3001\u8aac\u660e\u5909\u6570”Sepal.Length”\u3001”Sepal.Width”\u3001 “Petal.Length” “Petal.Width”\u304b\u3089\u5f93\u5c5e\u5909\u6570\u201dSpecies\u201d\u3092\u63a8\u5b9a\u3059\u308b\u305f\u3081\u306e\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u3092\u5b9f\u884c\u3059\u308b\u30b9\u30af\u30ea\u30d7\u30c8\u306f\uff1f\u5909\u6570dat\u306biris\u306e\u30c7\u30fc\u30bf\u304c\u683c\u7d0d\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/mark><\/p>\n\n\n\n

nnet\u30d1\u30c3\u30b1\u30fc\u30b8\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3001\u30d1\u30c3\u30b1\u30fc\u30b8\u306e\u8aad\u307f\u8fbc\u307f\u3001\u30c7\u30fc\u30bf\u306e\u6e96\u5099\u306e\u89e3\u8aac\u304c\u63d0\u793a\u3055\u308c\u307e\u3059\u3002<\/p>\n\n\n\n

# nnet\u30d1\u30c3\u30b1\u30fc\u30b8\u3092\u8aad\u307f\u8fbc\u3080 <\/p>\n\n\n\n

library(nnet)<\/p>\n\n\n\n

\u3053\u308c\u3089\u306b\u7d9a\u3044\u3066\u3001\u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u5206\u6790\u306e\u5b9f\u884c\u306e\u89e3\u8aac\u304c\u63d0\u793a\u3055\u308c\u3001multinom\u95a2\u6570\u3092\u4f7f\u3063\u3066\u5206\u6790\u3059\u308b\u30b9\u30af\u30ea\u30d7\u30c8\u304c\u63d0\u793a\u3055\u308c\u307e\u3059\u3002\u305d\u306e\u90e8\u5206\u3092\u4ee5\u4e0b\u306b\u793a\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

# \u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u30e2\u30c7\u30eb\u306e\u69cb\u7bc9<\/p>\n\n\n\n

# trace = FALSE \u306f\u3001\u6700\u9069\u5316\u306e\u904e\u7a0b\u3092\u8868\u793a\u3057\u306a\u3044\u3088\u3046\u306b\u3059\u308b\u305f\u3081\u306e\u3082\u306e\u3067\u3059<\/p>\n\n\n\n

model <- multinom(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, data = dat, trace = FALSE)<\/p>\n\n\n\n

# \u307e\u305f\u306f\u3001\u3059\u3079\u3066\u306e\u8aac\u660e\u5909\u6570\u3092\u4f7f\u7528\u3059\u308b\u5834\u5408<\/p>\n\n\n\n

# model <- multinom(Species ~ ., data = dat, trace = FALSE)<\/p>\n\n\n\n

\u305d\u3057\u3066\u3001\u7d50\u679c\u306e\u78ba\u8a8d\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u306f\uff1a<\/p>\n\n\n\n

# \u30e2\u30c7\u30eb\u306e\u8981\u7d04\u3092\u8868\u793a summary(model)<\/p>\n\n\n\n

\u3053\u308c\u3089\u3092R\u306e\u30a8\u30c7\u30a3\u30bf\u306b\u30b3\u30d4\u30fc\u30fb\u8cbc\u308a\u4ed8\u3051\u3057\u3066\u5b9f\u884c\u3057\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u3053\u306e\u3088\u3046\u306a\u7d50\u679c\u304c\u51fa\u529b\u3055\u308c\u307e\u3057\u305f\u3002\u5b9f\u306f\u3001Gemini\u304c\u51fa\u529b\u4f8b\u3068\u3057\u3066\u63d0\u793a\u3057\u305f\u5185\u5bb9\u3068\u306f\u7570\u306a\u3063\u3066\u3044\u307e\u3057\u305f\u3002\u3053\u306e\u7d50\u679c\u3092\u898b\u308b\u3068\u3001setosa\u3092\u53c2\u7167\u57fa\u6e96\u3068\u3057\u3066\u3001versicolor\u3001virginica\u306b\u5bfe\u3059\u308b\u4fc2\u6570\u304c\u8a08\u7b97\u3055\u308c\u3066\u3044\u307e\u3059\u3002Gemini\u306e\u51fa\u529b\u4f8b\u3067\u306f\u3001\u53c2\u7167\u57fa\u6e96\u304cvirginica\u306b\u306a\u3063\u3066\u304a\u308a\u3001setosa\u3001versicolor\u306b\u5bfe\u3059\u308b\u4fc2\u6570\u304c\u63d0\u793a\u3055\u308c\u3066\u3044\u307e\u3057\u305f\u3002\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u304c\u3053\u3061\u3089\u3067\u3059\u3002<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u305d\u308c\u3067\u306f\u3001\u672c\u5f53\u306b\u305d\u3046\u306a\u306e\u304b\u3001multinom\u95a2\u6570\u3067reference\u3092virginica\u306b\u5909\u3048\u308b\u306b\u306f\u3069\u3046\u3057\u305f\u3089\u3044\u3044\u304bGemini\u306b\u805e\u3044\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/p>\n\n\n\n

multinom\u95a2\u6570\u3067\u30ec\u30d5\u30a1\u30ec\u30f3\u30b9\u3092virginica\u306b\u3059\u308b\u306b\u306f\u3069\u3046\u3057\u307e\u3059\u304b\uff1f<\/mark><\/p>\n\n\n\n

\u30c7\u30d5\u30a9\u30eb\u30c8\u3067\u306f\u3001\u5f93\u5c5e\u5909\u6570\u306e\u4e26\u3073\u9806\u3067\u6700\u521d\u306e\u5909\u6570\u3092\u53c2\u7167\u57fa\u6e96reference\u306b\u3059\u308b\u3068\u3044\u3046\u8aac\u660e\u304c\u3055\u308c\u3001\u3082\u3057virginia\u306b\u5909\u66f4\u3059\u308b\u306e\u3067\u3042\u308c\u3070\u3001\u6b21\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u5b9f\u884c\u3059\u308b\u3068\u3044\u3046\u8aac\u660e\u3067\u3057\u305f\u3002\u5f93\u3063\u3066\u3001\u4e0a\u8a18\u306e\u51fa\u529b\u4f8b\u306f\u30c7\u30d5\u30a9\u30eb\u30c8\u3067\u306e\u89e3\u6790\u7d50\u679c\u3067\u306f\u306a\u3044\u3068\u3044\u3046\u3053\u3068\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n\n\n\n

\u53c2\u7167\u57fa\u6e96\u306e\u5909\u66f4\u304c\u3067\u304d\u308b\u304b\u3084\u3063\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

dat$Species <- relevel(dat$Species, ref = “virginica”)<\/p>\n\n\n\n

# \u30ec\u30d9\u30eb\u304c\u6b63\u3057\u304f\u5909\u66f4\u3055\u308c\u305f\u304b\u78ba\u8a8d <\/p>\n\n\n\n

levels(dat$Species)<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u3055\u3066\u3001\u4eca\u5ea6\u306f\u53c2\u7167\u57fa\u6e96virginia\u306b\u5bfe\u3059\u308b\u3001setosa\u3001versioclor\u306b\u5bfe\u3059\u308b\u4fc2\u6570\u306e\u5024\u304c\u51fa\u529b\u3055\u308c\u307e\u3057\u305f\u3002\u3057\u304b\u3057\u3001\u3053\u308c\u3082\u4e0a\u8a18\u306e\u51fa\u529b\u4f8b\u3068\u306f\u9055\u3046\u5024\u306b\u306a\u3063\u3066\u3044\u307e\u3059\u3002\u3064\u307e\u308a\u3001Gemini\u304c\u51fa\u529b\u4f8b\uff08\u4e00\u90e8\uff09\u3068\u3057\u3066\u63d0\u793a\u3057\u305f\u30c7\u30fc\u30bf\u306e\u7531\u6765\u306f\u5206\u304b\u308a\u307e\u305b\u3093\u304c\u3001\u3069\u3046\u3082\u9593\u9055\u3044\u306e\u3088\u3046\u3067\u3059\u3002<\/p>\n\n\n\n

\u3055\u3066\u3001\u5143\u306e\u30c7\u30fc\u30bf\u306e\u307e\u307e\u3001\u518d\u5ea6\u540c\u3058\u89e3\u6790\u3092\u5b9f\u884c\u3057\u3066\u304a\u304d\u307e\u3059\u3002\u305d\u3057\u3066\u3001Gemini\u304b\u3089\u306e\u56de\u7b54\u306b\u3042\u3063\u305f\u3001\u4fc2\u6570\u3068\u6a19\u6e96\u8aa4\u5dee\u3001p\u5024\uff08\u4e21\u5074\u691c\u5b9a\uff09\u306e\u5024\u3092\u51fa\u529b\u3057\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

# \u4fc2\u6570\u3068\u6a19\u6e96\u8aa4\u5dee\u3092\u53d6\u5f97 <\/p>\n\n\n\n

z <- summary(model)$coefficients \/ summary(model)$standard.errors <\/p>\n\n\n\n

# p\u5024\u3092\u8a08\u7b97 (\u4e21\u5074\u691c\u5b9a) <\/p>\n\n\n\n

p_values <- (1 – pnorm(abs(z), 0, 1)) * 2 print(p_values)<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u7d9a\u3044\u3066\u3001\u8aac\u660e\u5909\u6570\u306e\u5024\u304b\u3089\u3001Species\u306e\u63a8\u5b9aprediction\u3092\u3059\u308b\u306b\u306f\u3069\u3046\u3059\u308b\u304bGemini\u306b\u805e\u3044\u3066\u307f\u307e\u3057\u3087\u3046\u3002\u307e\u305a\u3001\u5206\u6790\u306e\u5143\u306e\u30c7\u30fc\u30bf\u3001\u3064\u307e\u308a\u8a13\u7df4\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u9069\u7528\u3057\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

dat\u306e4\u3064\u306e\u8aac\u660e\u5909\u6570\u306e\u5024\u304b\u3089Species\u3092\u63a8\u5b9a\u3059\u308b\u306b\u306f\u3069\u3046\u3057\u307e\u3059\u304b\uff1f<\/mark><\/p>\n\n\n\n

predict\u95a2\u6570\u3092\u4f7f\u7528\u3059\u308b\u3053\u3068\u3068\u3001\u8a13\u7df4\u30c7\u30fc\u30bf\u306e\u5834\u5408\u306e\u63a8\u5b9a\u7d50\u679c\u3001\u6b63\u89e3\u7387\u306e\u8a08\u7b97\u3001\u63a8\u5b9a\u78ba\u7387\u306e\u8a08\u7b97\u306e\u305f\u3081\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u304c\u63d0\u793a\u3055\u308c\u307e\u3057\u305f\u3002Species\u3001\u3064\u307e\u308a\u30a2\u30e4\u30e1\u306e\u540d\u79f0\u3092\u30af\u30e9\u30b9\u3068\u547c\u3093\u3067\u3044\u307e\u3059\u3002<\/p>\n\n\n\n

# \u8a13\u7df4\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u63a8\u5b9a\u3055\u308c\u305f\u30af\u30e9\u30b9<\/p>\n\n\n\n

predicted_class <- predict(model, newdata = dat, type = “class”)<\/p>\n\n\n\n

print(head(predicted_class))<\/p>\n\n\n\n

# \u5b9f\u969b\u306eSpecies\u3068\u6bd4\u8f03\u3057\u3066\u3001\u3069\u306e\u7a0b\u5ea6\u5408\u3063\u3066\u3044\u308b\u304b\u78ba\u8a8d\u3059\u308b\uff08\u6df7\u540c\u884c\u5217\uff09<\/p>\n\n\n\n

confusion_matrix <- table(Actual = dat$Species, Predicted = predicted_class)<\/p>\n\n\n\n

print(confusion_matrix)<\/p>\n\n\n\n

# \u6b63\u89e3\u7387\u306e\u8a08\u7b97<\/p>\n\n\n\n

accuracy <- sum(diag(confusion_matrix)) \/ sum(confusion_matrix)<\/p>\n\n\n\n

cat(“\u30e2\u30c7\u30eb\u306e\u6b63\u89e3\u7387:”, round(accuracy * 100, 2), “%\\n”)<\/p>\n\n\n\n

# \u8a13\u7df4\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u5404\u30af\u30e9\u30b9\u306e\u63a8\u5b9a\u78ba\u7387<\/p>\n\n\n\n

predicted_probs <- predict(model, newdata = dat, type = “probs”)<\/p>\n\n\n\n

print(head(predicted_probs))<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

<\/p>\n\n\n\n

iris\u306e\u30c7\u30fc\u30bf\u306f\u3001nrow(dat)\u3067150\u3001\u3064\u307e\u308a150\u500b\u4f53\u306e\u30a2\u30a4\u30ea\u30b9\u306e\u30c7\u30fc\u30bf\u3092\u542b\u3093\u3067\u3044\u307e\u3059\u3002\u3053\u3053\u3067\u306e\u51fa\u529b\u306fhead()\u95a2\u6570\u3067\u6700\u521d\u306e6\u500b\u3060\u3051\u306b\u306a\u3063\u3066\u3044\u307e\u3059\u3002confusion_matrix\u3064\u307e\u308a\u6df7\u540c\u884c\u5217\u306e\u7d50\u679c\u3092\u307f\u308b\u3068\u3001setosa50\u500b\u4f53\u306f\u3059\u3079\u3066setosa\u3068\u3057\u3066\u63a8\u5b9a\u3055\u308c\u3001versicolor\u306f50\u500b\u4f53\u306e\u518549\u500b\u4f53\u304c\u6b63\u3057\u304fversicolor\u3068\u3057\u3066\u63a8\u5b9a\u3055\u308c\u3001virginica1\u306f50\u500b\u4f53\u306e\u518549\u500b\u4f53\u304c\u6b63\u3057\u304fvirginica\u3068\u3057\u3066\u5206\u985e\u3055\u308c\u3066\u3044\u308b\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3059\u3002\u6b63\u89e3\u7387\u306e\u8a08\u7b97\u3067\u306f98.67\uff05\u3067\u3057\u305f\u3002\u975e\u5e38\u306b\u9ad8\u3044\u7cbe\u5ea6\u3067\u3059\u3002<\/p>\n\n\n\n

\u6700\u5f8c\u306e\u3001\u8a13\u7df4\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u5404\u30af\u30e9\u30b9\u306e\u63a8\u5b9a\u78ba\u7387\u3092\u898b\u308b\u3068\u500b\u4f531\u306fsetosa\u306e\u78ba\u7387\u304c1.0\u3067versicolor\u3001verginica\u306e\u78ba\u7387\u306f\u307b\u307c0\u3068\u3044\u3046\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3059\u3002\u4ee5\u964d\u500b\u4f536\u307e\u3067\u306e\u7d50\u679c\u304c\u51fa\u529b\u3055\u308c\u3066\u3044\u307e\u3059\u304c\u3001print(predicted_probs)\u3067\u3001150\u500b\u4f53\u3059\u3079\u3066\u306e\u7d50\u679c\u304c\u51fa\u529b\u3055\u308c\u307e\u3059\u3002<\/p>\n\n\n\n

\u3055\u3066\u3001Gemini\u306e\u56de\u7b54\u306b\u306f\u3001\u65b0\u3057\u3044\u30c7\u30fc\u30bf\u30dd\u30a4\u30f3\u30c8\u306b\u3064\u3044\u3066Species\u3092\u63a8\u5b9a\u3059\u308b\u65b9\u6cd5\u3001\u30b9\u30af\u30ea\u30d7\u30c8\u306b\u3064\u3044\u3066\u3082\u8aac\u660e\u304c\u3042\u308a\u307e\u3059\u3002\u4ee5\u4e0b\u304c\u305d\u306e\u90e8\u5206\u3067\u3059\u3002\u81e8\u5e8a\u306e\u5834\u5408\u3067\u3042\u308c\u3070\u3001\u500b\u3005\u306e\u5bfe\u8c61\u8005\u306e\u30c7\u30fc\u30bf\u304b\u3089\u3001\u8a3a\u65ad\u3001\u3042\u308b\u3044\u306f\u75c5\u614b\u3092\u63a8\u5b9a\u3059\u308b\u3053\u3068\u306b\u76f8\u5f53\u3057\u307e\u3059\u3002\u3053\u3053\u3067\u306f\u3001\uff13\u3064\u306e\u500b\u4f53\u306e\u30c7\u30fc\u30bf\u3092\u4f5c\u6210\u3057\u3066\u3001\u63a8\u5b9a\u3092\u884c\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n\n\n\n

——–
2. \u65b0\u3057\u3044\u30c7\u30fc\u30bf\u30dd\u30a4\u30f3\u30c8\u306b\u3064\u3044\u3066Species\u3092\u63a8\u5b9a\u3059\u308b<\/strong><\/p>\n\n\n\n

\u3053\u308c\u306f\u3001\u30e2\u30c7\u30eb\u3092\u672a\u77e5\u306e\u30c7\u30fc\u30bf\u306b\u9069\u7528\u3059\u308b\u5834\u5408\u306e\u4e00\u822c\u7684\u306a\u4f7f\u7528\u65b9\u6cd5\u3067\u3059\u3002\u65b0\u3057\u3044\u30c7\u30fc\u30bf\u306f\u3001\u30e2\u30c7\u30eb\u69cb\u7bc9\u6642\u306b\u4f7f\u7528\u3057\u305f\u8aac\u660e\u5909\u6570\u3068\u540c\u3058\u30ab\u30e9\u30e0\u540d\u3068\u30c7\u30fc\u30bf\u578b\u3092\u6301\u3064\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u3067\u3042\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n\n\n\n

\u4f8b\u3068\u3057\u3066\u3001\u65b0\u3057\u3044\u30a2\u30e4\u30e1\u306e\u6e2c\u5b9a\u5024\u3092\u4f7f\u3063\u3066Species\u3092\u63a8\u5b9a\u3057\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/p>\n\n\n\n

R<\/p>\n\n\n\n

# \u65b0\u3057\u3044\u30c7\u30fc\u30bf\u30dd\u30a4\u30f3\u30c8\u3092\u4f5c\u6210<\/p>\n\n\n\n

# \u30ab\u30e9\u30e0\u540d\u306f\u30e2\u30c7\u30eb\u69cb\u7bc9\u6642\u306b\u4f7f\u7528\u3057\u305f\u8aac\u660e\u5909\u6570\u3068\u5b8c\u5168\u306b\u4e00\u81f4\u3055\u305b\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059<\/p>\n\n\n\n

new_iris_data <- data.frame(<\/p>\n\n\n\n

  Sepal.Length = c(5.1, 6.0, 7.0),<\/p>\n\n\n\n

  Sepal.Width = c(3.5, 2.7, 3.2),<\/p>\n\n\n\n

  Petal.Length = c(1.4, 4.5, 6.0),<\/p>\n\n\n\n

  Petal.Width = c(0.2, 1.5, 2.0)<\/p>\n\n\n\n

)<\/p>\n\n\n\n

print(new_iris_data)<\/p>\n\n\n\n

# \u65b0\u3057\u3044\u30c7\u30fc\u30bf\u30dd\u30a4\u30f3\u30c8\u306b\u5bfe\u3059\u308b\u63a8\u5b9a\u3055\u308c\u305f\u30af\u30e9\u30b9<\/p>\n\n\n\n

predicted_new_class <- predict(model, newdata = new_iris_data, type = “class”)<\/p>\n\n\n\n

cat(“\\n\u65b0\u3057\u3044\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u63a8\u5b9a\u3055\u308c\u305fSpecies:\\n”)<\/p>\n\n\n\n

print(predicted_new_class)<\/p>\n\n\n\n

# \u65b0\u3057\u3044\u30c7\u30fc\u30bf\u30dd\u30a4\u30f3\u30c8\u306b\u5bfe\u3059\u308b\u5404Species\u30ab\u30c6\u30b4\u30ea\u306e\u63a8\u5b9a\u78ba\u7387<\/p>\n\n\n\n

predicted_new_probs <- predict(model, newdata = new_iris_data, type = “probs”)<\/p>\n\n\n\n

cat(“\\n\u65b0\u3057\u3044\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u5404Species\u306e\u63a8\u5b9a\u78ba\u7387:\\n”)<\/p>\n\n\n\n

print(predicted_new_probs)<\/p>\n\n\n\n

\u3053\u306e\u3088\u3046\u306b\u3001predict()\u95a2\u6570\u3092\u4f7f\u3046\u3053\u3068\u3067\u3001\u5b66\u7fd2\u6e08\u307f\u30e2\u30c7\u30eb\u3092\u7528\u3044\u3066\u8aac\u660e\u5909\u6570\u306e\u5024\u304b\u3089\u5f93\u5c5e\u5909\u6570Species\u3092\u63a8\u5b9a\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002<\/p>\n\n\n\n

—————<\/p>\n\n\n\n

\u3053\u308c\u3089\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u3092R\u3067\u5b9f\u884c\u3057\u3066\u307f\u307e\u3059\u3002<\/p>\n\n\n\n

\"\"\/<\/figure>\n\n\n\n

\u5f93\u3063\u3066\u3001\u500b\u4f531\u306fsetosa\u3001\u500b\u4f532\u306fversicolor\u3001\u500b\u4f533\u306fvirginica\u3068\u63a8\u5b9a\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n\n\n\n

\u3055\u3066iris\u306e\u30c7\u30fc\u30bf\u306f\u8aac\u660e\u5909\u6570\u304c\u9023\u7d9a\u5909\u6570\u306e\u307f\u3067\u3057\u305f\u3002\u8aac\u660e\u5909\u6570\u306b\u540d\u7fa9\u5909\u6570\u304c\u3042\u308b\u5834\u5408\u3001\u3069\u306e\u3088\u3046\u306b\u6271\u3046\u306e\u304b\u3001\u7279\u306b\u53c2\u7167\u57fa\u6e96\u306e\u8a2d\u5b9a\u306b\u3064\u3044\u3066\u805e\u3044\u3066\u307f\u307e\u3057\u3087\u3046\u3002<\/p>\n\n\n\n

\u8aac\u660e\u5909\u6570\u306b\u540d\u7fa9\u5909\u6570\u304c\u3042\u308b\u5834\u5408\u3001\u4f8b\u3048\u3070\u3001x1: A, B, C, D\uff1bx2: X, Y, Z\u306e\u3088\u3046\u306a\u5834\u5408\u3001multinom\u95a2\u6570\u3092\u4f7f\u3046\u5834\u5408\u306b\u3001\u53c2\u7167\u57fa\u6e96\u306e\u8a2d\u5b9a\u306f\uff1f<\/mark><\/p>\n\n\n\n

R\u306emultinom\u95a2\u6570\u3067\u3001\u8aac\u660e\u5909\u6570\u306b\u540d\u7fa9\u5909\u6570\uff08\u30ab\u30c6\u30b4\u30ea\u30ab\u30eb\u5909\u6570\uff09\u304c\u3042\u308b\u5834\u5408\u3001\u305d\u306e\u8aac\u660e\u5909\u6570\u306e\u53c2\u7167\u57fa\u6e96\uff08\u30ec\u30d5\u30a1\u30ec\u30f3\u30b9\u30ec\u30d9\u30eb\uff09\u3092\u8a2d\u5b9a\u3059\u308b\u65b9\u6cd5\u306f\u3001\u5f93\u5c5e\u5909\u6570\u306e\u5834\u5408\u3068\u5168\u304f\u540c\u3058\u8003\u3048\u65b9\u3067\u3059\u3002<\/p>\n\n\n\n

R\u306e\u7d71\u8a08\u30e2\u30c7\u30ea\u30f3\u30b0\u95a2\u6570\uff08lm, glm, multinom\u306a\u3069\uff09\u306f\u3001\u56e0\u5b50\uff08factor\uff09\u578b\u5909\u6570<\/strong>\u3092\u81ea\u52d5\u7684\u306b\u30c0\u30df\u30fc\u5909\u6570\u306b\u5909\u63db\u3057\u3066\u56de\u5e30\u5206\u6790\u306b\u7d44\u307f\u8fbc\u307f\u307e\u3059\u3002\u3053\u306e\u969b\u3001\u30c7\u30d5\u30a9\u30eb\u30c8\u3067\u306f\u305d\u306e\u56e0\u5b50\u306e\u6700\u521d\u306e\u30ec\u30d9\u30eb<\/strong>\u304c\u53c2\u7167\u57fa\u6e96\u3068\u306a\u308a\u3001\u305d\u306e\u30ec\u30d9\u30eb\u306b\u5bfe\u5fdc\u3059\u308b\u30c0\u30df\u30fc\u5909\u6570\u306f\u30e2\u30c7\u30eb\u304b\u3089\u9664\u5916\uff08\u4fc2\u6570\u304c0\u3068\u3057\u3066\u6271\u308f\u308c\u308b\uff09\u3055\u308c\u307e\u3059\u3002\u4ed6\u306e\u30ec\u30d9\u30eb\u306e\u4fc2\u6570\u306f\u3001\u3053\u306e\u53c2\u7167\u57fa\u6e96\u30ec\u30d9\u30eb\u3068\u6bd4\u8f03\u3057\u3066\u3069\u306e\u304f\u3089\u3044\u7570\u306a\u308b\u304b\u3092\u793a\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

\u3057\u305f\u304c\u3063\u3066\u3001\u8aac\u660e\u5909\u6570x1\u3084x2\u306e\u53c2\u7167\u57fa\u6e96\u3092\u8a2d\u5b9a\u3057\u305f\u3044\u5834\u5408\u306f\u3001\u305d\u308c\u305e\u308c\u306e\u5909\u6570\u3092factor\u578b\u3067\u3042\u308b\u3053\u3068\u3092\u78ba\u8a8d\u3057\u3001relevel()\u95a2\u6570\u3092\u4f7f\u3063\u3066\u53c2\u7167\u57fa\u6e96\u306b\u3057\u305f\u3044\u30ec\u30d9\u30eb\u3092\u6700\u521d\u306e\u30ec\u30d9\u30eb\u306b\u8a2d\u5b9a\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

\u4ee5\u4e0b\u306b\u5177\u4f53\u7684\u306a\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u793a\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

1. \u30b5\u30f3\u30d7\u30eb\u30c7\u30fc\u30bf\u306e\u4f5c\u6210<\/strong><\/p>\n\n\n\n

\u307e\u305a\u3001\u3054\u63d0\u793a\u306ex1 (A, B, C, D) \u3068 x2 (X, Y, Z) \u306e\u3088\u3046\u306a\u540d\u7fa9\u5909\u6570\u3092\u542b\u3080\u30c0\u30df\u30fc\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u4f5c\u6210\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

R<\/p>\n\n\n\n

# \u30c7\u30fc\u30bf\u3092\u30b7\u30df\u30e5\u30ec\u30fc\u30c8<\/p>\n\n\n\n

set.seed(123)<\/p>\n\n\n\n

n <- 100 # \u30b5\u30f3\u30d7\u30eb\u30b5\u30a4\u30ba<\/p>\n\n\n\n

# \u8aac\u660e\u5909\u6570 x1 (A, B, C, D)<\/p>\n\n\n\n

x1_levels <- c(“A”, “B”, “C”, “D”)<\/p>\n\n\n\n

x1 <- sample(x1_levels, n, replace = TRUE, prob = c(0.25, 0.25, 0.25, 0.25))<\/p>\n\n\n\n

# \u8aac\u660e\u5909\u6570 x2 (X, Y, Z)<\/p>\n\n\n\n

x2_levels <- c(“X”, “Y”, “Z”)<\/p>\n\n\n\n

x2 <- sample(x2_levels, n, replace = TRUE, prob = c(0.33, 0.33, 0.34))<\/p>\n\n\n\n

# \u5f93\u5c5e\u5909\u6570 Species (iris\u30c7\u30fc\u30bf\u3068\u540c\u69d8\u306b3\u30ab\u30c6\u30b4\u30ea\u3092\u60f3\u5b9a)<\/p>\n\n\n\n

species_levels <- c(“setosa”, “versicolor”, “virginica”)<\/p>\n\n\n\n

species <- sample(species_levels, n, replace = TRUE) # \u30e9\u30f3\u30c0\u30e0\u306aSpecies\u3092\u751f\u6210<\/p>\n\n\n\n

# \u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u3092\u4f5c\u6210<\/p>\n\n\n\n

my_dat <- data.frame(<\/p>\n\n\n\n

  Species = as.factor(species), # \u5f93\u5c5e\u5909\u6570\u306ffactor\u578b<\/p>\n\n\n\n

  x1 = as.factor(x1),            # \u8aac\u660e\u5909\u6570x1\u306ffactor\u578b<\/p>\n\n\n\n

  x2 = as.factor(x2)             # \u8aac\u660e\u5909\u6570x2\u306ffactor\u578b<\/p>\n\n\n\n

)<\/p>\n\n\n\n

# \u30c7\u30fc\u30bf\u306e\u69cb\u9020\u3092\u78ba\u8a8d<\/p>\n\n\n\n

str(my_dat)<\/p>\n\n\n\n

2. \u8aac\u660e\u5909\u6570\u306e\u53c2\u7167\u57fa\u6e96\u306e\u8a2d\u5b9a<\/strong><\/p>\n\n\n\n

relevel()\u95a2\u6570\u3092\u4f7f\u3063\u3066\u3001\u5404\u8aac\u660e\u5909\u6570\u306e\u53c2\u7167\u57fa\u6e96\u3092\u5909\u66f4\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

x1\u306e\u53c2\u7167\u57fa\u6e96\u3092<\/strong>B\u306b\u3059\u308b\u5834\u5408:
<\/strong>R<\/p>\n\n\n\n

my_dat$x1 <- relevel(my_dat$x1, ref = “B”)<\/p>\n\n\n\n

levels(my_dat$x1) # \u78ba\u8a8d: “B” \u304c\u6700\u521d\u306b\u306a\u3063\u3066\u3044\u308b\u306f\u305a<\/p>\n\n\n\n

x2\u306e\u53c2\u7167\u57fa\u6e96\u3092<\/strong>Z\u306b\u3059\u308b\u5834\u5408:
<\/strong>R<\/p>\n\n\n\n

my_dat$x2 <- relevel(my_dat$x2, ref = “Z”)<\/p>\n\n\n\n

levels(my_dat$x2) # \u78ba\u8a8d: “Z” \u304c\u6700\u521d\u306b\u306a\u3063\u3066\u3044\u308b\u306f\u305a<\/p>\n\n\n\n

3. <\/strong>multinom\u95a2\u6570\u306e\u5b9f\u884c<\/strong><\/p>\n\n\n\n

\u53c2\u7167\u57fa\u6e96\u3092\u8a2d\u5b9a\u3057\u305f\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0my_dat\u3092\u4f7f\u3063\u3066multinom\u95a2\u6570\u3092\u5b9f\u884c\u3057\u307e\u3059\u3002<\/p>\n\n\n\n

R<\/p>\n\n\n\n

library(nnet)<\/p>\n\n\n\n

# (\u30aa\u30d7\u30b7\u30e7\u30f3) \u5f93\u5c5e\u5909\u6570Species\u306e\u53c2\u7167\u57fa\u6e96\u3082\u8a2d\u5b9a\u3059\u308b\u4f8b (\u3082\u3057\u5fc5\u8981\u306a\u3089)<\/p>\n\n\n\n

# my_dat$Species <- relevel(my_dat$Species, ref = “virginica”)<\/p>\n\n\n\n

# \u591a\u9805\u30ed\u30b8\u30b9\u30c6\u30a3\u30c3\u30af\u56de\u5e30\u30e2\u30c7\u30eb\u306e\u69cb\u7bc9<\/p>\n\n\n\n

# trace = FALSE \u306f\u3001\u6700\u9069\u5316\u306e\u904e\u7a0b\u3092\u8868\u793a\u3057\u306a\u3044\u3088\u3046\u306b\u3059\u308b\u305f\u3081\u306e\u3082\u306e\u3067\u3059<\/p>\n\n\n\n

model_nominal <- multinom(Species ~ x1 + x2, data = my_dat, trace = FALSE)<\/p>\n\n\n\n

# \u30e2\u30c7\u30eb\u306e\u8981\u7d04\u3092\u8868\u793a<\/p>\n\n\n\n

summary(model_nominal)<\/p>\n\n\n\n

\u51fa\u529b\u306e\u89e3\u91c8<\/strong><\/p>\n\n\n\n

summary(model_nominal)\u306eCoefficients\u30bb\u30af\u30b7\u30e7\u30f3\u3092\u898b\u308b\u3068\u3001x1\u3068x2\u306e\u4fc2\u6570\u304c\u3069\u306e\u3088\u3046\u306b\u8868\u793a\u3055\u308c\u308b\u304b\u3092\u78ba\u8a8d\u3067\u304d\u307e\u3059\u3002<\/p>\n\n\n\n

    \n
  • x1\u306b\u3064\u3044\u3066:
    x1B\u306e\u3088\u3046\u306a\u4fc2\u6570\u306f\u8868\u793a\u3055\u308c\u305a\u3001x1A\u3001x1C\u3001x1D\u306e\u3088\u3046\u306a\u4fc2\u6570\u304c\u8868\u793a\u3055\u308c\u307e\u3059\u3002\u3053\u308c\u306f\u3001\u305d\u308c\u305e\u308c\u306e\u30ec\u30d9\u30eb\uff08A, C, D\uff09\u304c\u53c2\u7167\u57fa\u6e96B\u3068\u6bd4\u8f03\u3057\u3066\u3069\u3046\u3067\u3042\u308b\u304b\u3092\u793a\u3057\u307e\u3059\u3002
    <\/li>\n\n\n\n
  • x2\u306b\u3064\u3044\u3066:
    x2Z\u306e\u3088\u3046\u306a\u4fc2\u6570\u306f\u8868\u793a\u3055\u308c\u305a\u3001x2X\u3001x2Y\u306e\u3088\u3046\u306a\u4fc2\u6570\u304c\u8868\u793a\u3055\u308c\u307e\u3059\u3002\u3053\u308c\u306f\u3001\u305d\u308c\u305e\u308c\u306e\u30ec\u30d9\u30eb\uff08X, Y\uff09\u304c\u53c2\u7167\u57fa\u6e96Z\u3068\u6bd4\u8f03\u3057\u3066\u3069\u3046\u3067\u3042\u308b\u304b\u3092\u793a\u3057\u307e\u3059\u3002
    <\/li>\n<\/ul>\n\n\n\n

    \u91cd\u8981\u306a\u6ce8\u610f\u70b9:<\/strong><\/p>\n\n\n\n

      \n
    • factor\u578b\u3067\u3042\u308b\u3053\u3068\u306e\u78ba\u8a8d:<\/strong> \u8aac\u660e\u5909\u6570\u304cfactor\u578b\u3067\u3042\u308b\u3053\u3068\u3092\u5e38\u306b\u78ba\u8a8d\u3057\u3066\u304f\u3060\u3055\u3044\u3002\u6587\u5b57\u5217\u578b\u306e\u307e\u307e\u3067\u306fR\u306f\u81ea\u52d5\u7684\u306b\u30c0\u30df\u30fc\u5909\u6570\u5316\u3057\u307e\u305b\u3093\u3002<\/li>\n\n\n\n
    • \u30c7\u30d5\u30a9\u30eb\u30c8\u306e\u52d5\u4f5c:<\/strong> relevel()\u3092\u4f7f\u7528\u3057\u306a\u3044\u5834\u5408\u3001R\u306f\u30a2\u30eb\u30d5\u30a1\u30d9\u30c3\u30c8\u9806\uff08\u307e\u305f\u306f\u6570\u5024\u9806\uff09\u3067\u6700\u521d\u306e\u30ec\u30d9\u30eb\u3092\u53c2\u7167\u57fa\u6e96\u3068\u3057\u307e\u3059\u3002<\/li>\n\n\n\n
    • \u4ed6\u306e\u53c2\u7167\u57fa\u6e96\u8a2d\u5b9a\u65b9\u6cd5:<\/strong> C()\u95a2\u6570\u3068contr.treatments()\u3092\u7d44\u307f\u5408\u308f\u305b\u3066\u3001\u30e2\u30c7\u30eb\u306e\u5f0f\u5185\u3067\u4e00\u6642\u7684\u306b\u53c2\u7167\u57fa\u6e96\u3092\u8a2d\u5b9a\u3059\u308b\u3053\u3068\u3082\u53ef\u80fd\u3067\u3059\u304c\u3001relevel()\u3067\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u81ea\u4f53\u3092\u5909\u66f4\u3059\u308b\u65b9\u304c\u3001\u30b3\u30fc\u30c9\u304c\u8aad\u307f\u3084\u3059\u304f\u306a\u308b\u3053\u3068\u304c\u591a\u3044\u3067\u3059\u3002 \u4f8b: multinom(Species ~ C(x1, contr.treatment(nlevels(x1), base=2)) + x2, data = my_dat) (\u3053\u306e\u4f8b\u3067\u306fx1\u306e2\u756a\u76ee\u306e\u30ec\u30d9\u30eb\u3092\u57fa\u6e96\u306b\u3059\u308b)<\/li>\n<\/ul>\n\n\n\n

      \u3053\u306e\u3088\u3046\u306b\u3001\u8aac\u660e\u5909\u6570\u306b\u540d\u7fa9\u5909\u6570\u304c\u3042\u308b\u5834\u5408\u3067\u3082\u3001relevel()\u95a2\u6570\u3092\u4f7f\u7528\u3057\u3066\u53c2\u7167\u57fa\u6e96\u3092\u660e\u793a\u7684\u306b\u8a2d\u5b9a\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002
      —————<\/p>\n\n\n\n

      \u3053\u306e\u89e3\u8aac\u3067\u5206\u304b\u308b\u3088\u3046\u306b\u3001\u8aac\u660e\u5909\u6570\u304c\u540d\u7fa9\u5909\u6570\u306e\u5834\u5408\u306f\u3001\u53c2\u7167\u57fa\u6e96\u306frelevel\u3067\u8a2d\u5b9a\u3067\u304d\u308b\u3053\u3068\u3001\u7d50\u679c\u3067\u5f97\u3089\u308c\u308b\u4fc2\u6570\u306f\u305d\u308c\u305e\u308c\u306e\u540d\u7fa9\u5909\u6570\u306e\u5024\u306b\u5bfe\u3059\u308b\u5024\u3067\u3042\u308b\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3059\u3002<\/p>\n\n\n\n

      \u3055\u3089\u306b\u3001\u540d\u7fa9\u5909\u6570\u306e\u5834\u5408\u3001Multi-dichotomous variable\u306b\u3042\u3089\u304b\u3058\u3081\u624b\u4f5c\u696d\u3067\u5909\u63db\u3059\u308b\u5fc5\u8981\u304c\u3042\u308b\u304b\u805e\u3044\u305f\u3068\u3053\u308d\u3001\u305d\u306e\u5fc5\u8981\u306f\u306a\u3044\u3068\u3044\u3046\u56de\u7b54\u3067\u3057\u305f\u3002<\/p>\n\n\n\n

      \u8aac\u660e\u5909\u6570\u304c\u540d\u7fa9\u5909\u6570\u306e\u5834\u5408\u3001multi-dichootmous variable\u306b\u3042\u3089\u304b\u3058\u3081\u5909\u63db\u3059\u308b\u5fc5\u8981\u306f\u3042\u308a\u307e\u3059\u304b\uff1f<\/mark><\/p>\n\n\n\n

      \u3044\u3044\u3048\u3001R\u306emultinom\u95a2\u6570\uff08\u304a\u3088\u3073\u591a\u304f\u306e\u7d71\u8a08\u30e2\u30c7\u30ea\u30f3\u30b0\u95a2\u6570\u3001\u4f8b\uff1alm, glm\uff09\u3092\u4f7f\u7528\u3059\u308b\u5834\u5408\u3001\u8aac\u660e\u5909\u6570\u3067\u3042\u308b\u540d\u7fa9\u5909\u6570\uff08\u30ab\u30c6\u30b4\u30ea\u30ab\u30eb\u5909\u6570\uff09\u3092\u4e8b\u524d\u306b\u624b\u52d5\u3067multi-dichotomous variable\uff08\u591a\u5024\u4e8c\u9805\u5909\u6570\u3001\u3064\u307e\u308a\u30c0\u30df\u30fc\u5909\u6570\uff09\u306b\u5909\u63db\u3059\u308b\u5fc5\u8981\u306f\u3042\u308a\u307e\u305b\u3093\u3002<\/strong><\/p>\n\n\n\n

      R\u306f\u3001\u30c7\u30fc\u30bf\u30d5\u30ec\u30fc\u30e0\u5185\u306e\u5909\u6570\u304cfactor\u578b\u3068\u3057\u3066\u5b9a\u7fa9\u3055\u308c\u3066\u3044\u308c\u3070\u3001multinom\u95a2\u6570\u306b\u6e21\u3055\u308c\u305f\u3068\u304d\u306b\u3001\u5185\u90e8\u7684\u306b\u81ea\u52d5\u3067\u30c0\u30df\u30fc\u5909\u6570\u306b\u5909\u63db\u3057\u3066\u304f\u308c\u307e\u3059\u3002\u3053\u308c\u304cR\u306e\u30e2\u30c7\u30ea\u30f3\u30b0\u95a2\u6570\u306e\u4fbf\u5229\u306a\u70b9\u306e1\u3064\u3067\u3059\u3002<\/p>\n\n\n\n

      R\u304c\u81ea\u52d5\u7684\u306b\u884c\u3046\u3053\u3068:<\/strong><\/p>\n\n\n\n

        \n
      1. factor\u578b\u5909\u6570\u306e\u691c\u51fa<\/strong>: \u30e2\u30c7\u30eb\u5f0f\uff08\u4f8b: Species ~ x1 + x2\uff09\u306e\u4e2d\u306bfactor\u578b\u306e\u5909\u6570\u304c\u3042\u308b\u3068\u3001R\u306f\u305d\u308c\u3092\u81ea\u52d5\u7684\u306b\u8a8d\u8b58\u3057\u307e\u3059\u3002<\/li>\n\n\n\n
      2. \u30c0\u30df\u30fc\u5909\u6570\u306e\u751f\u6210<\/strong>: \u4f8b\u3048\u3070\u3001x1\u304cA, B, C, D\u306e4\u3064\u306e\u30ec\u30d9\u30eb\u3092\u6301\u3064factor\u578b\u306e\u5834\u5408\u3001R\u306f\u81ea\u52d5\u7684\u306b3\u3064\u306e\u30c0\u30df\u30fc\u5909\u6570\u3092\u751f\u6210\u3057\u307e\u3059\u3002\n