{"id":4731,"date":"2026-08-15T09:27:45","date_gmt":"2026-08-15T00:27:45","guid":{"rendered":"https:\/\/best-biostatistics.com\/toukei-er\/?p=4731"},"modified":"2026-08-15T09:27:45","modified_gmt":"2026-08-15T00:27:45","slug":"how-to-analyze-30-covariates-and-non-linear-risks-solutions-via-random-survival-forest-rsf-and-the-randomforestsrc-package","status":"publish","type":"post","link":"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-analyze-30-covariates-and-non-linear-risks-solutions-via-random-survival-forest-rsf-and-the-randomforestsrc-package\/","title":{"rendered":"30\u9805\u76ee\u4ee5\u4e0a\u306e\u80cc\u666f\u56e0\u5b50\u3068\u975e\u7dda\u5f62\u30ea\u30b9\u30af\u3092\u3069\u3046\u89e3\u6790\u3059\u3079\u304d\u304b\uff1fRandom Survival Forest\uff08RSF\uff09\u3068randomForestSRC\u30d1\u30c3\u30b1\u30fc\u30b8\u306b\u3088\u308b\u89e3\u6c7a\u6cd5"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">\u300c\u5148\u751f\uff01\u591a\u5909\u91cfCox\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u30e2\u30c7\u30eb\u3092\u7d44\u3082\u3046\u3068\u3057\u305f\u3089\u3001\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\uff08PH\u4eee\u5b9a\uff09\u304c\u8907\u6570\u306e\u5909\u6570\u3067\u5d29\u308c\u3066\u3057\u307e\u3044\u3001\u5358\u4e00\u306e\u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u51fa\u305b\u306a\u304f\u306a\u308a\u307e\u3057\u305f\uff01\u304a\u307e\u3051\u306b\u80cc\u666f\u56e0\u5b50\u304c30\u9805\u76ee\u4ee5\u4e0a\u3042\u3063\u3066\u3001\u30b9\u30c6\u30c3\u30d7\u30ef\u30a4\u30ba\u6cd5\u3092\u5165\u308c\u3066\u3082\u30e2\u30c7\u30eb\u304c\u904e\u5b66\u7fd2\u3067\u5d29\u58ca\u3057\u3066\u3057\u307e\u3044\u307e\u3059\u2026\u2026\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u591a\u5909\u91cfCox\u56de\u5e30\u3092\u7528\u3044\u3066\u4e88\u5f8c\u4e88\u6e2c\u3084\u5371\u967a\u56e0\u5b50\u63a2\u7d22\u3092\u884c\u3063\u3066\u3044\u308b\u3068\u3001\u3053\u3046\u3057\u305f\u300c\u624b\u8a70\u307e\u308a\u300d\u306b\u76f4\u9762\u3059\u308b\u3053\u3068\u304c\u5c11\u306a\u304f\u306a\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u81e8\u5e8a\u30c7\u30fc\u30bf\u3084\u30aa\u30df\u30af\u30b9\u30c7\u30fc\u30bf\u304c\u9ad8\u6b21\u5143\u5316\u30fb\u591a\u9805\u76ee\u5316\u3059\u308b\u73fe\u4ee3\u306b\u304a\u3044\u3066\u3001\u5f93\u6765\u306eCox\u56de\u5e30\u304c\u524d\u63d0\u3068\u3059\u308b\u300c\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\u300d\u3084\u300c\u5358\u4e00\u306e\u7dda\u5f62\u95a2\u4fc2\u300d\u306b\u30c7\u30fc\u30bf\u3092\u7121\u7406\u3084\u308a\u5f53\u3066\u306f\u3081\u308b\u3053\u3068\u306b\u306f\u9650\u754c\u304c\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3057\u304b\u3057\u3001\u300c\u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u51fa\u306a\u3044\u304b\u3089\u6a5f\u68b0\u5b66\u7fd2\u306f\u533b\u5b66\u8ad6\u6587\u3067\u4f7f\u3048\u306a\u3044\u300d\u3068\u601d\u3044\u8fbc\u3093\u3067\u8ae6\u3081\u308b\u5fc5\u8981\u306f\u5168\u304f\u306a\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u53f3\u5074\u6253\u5207\u308a\uff08Censoring\uff09\u306b\u5bfe\u5fdc\u3057\u305f\u6a5f\u68b0\u5b66\u7fd2\u624b\u6cd5\u3067\u3042\u308b <strong>Random Survival Forest\uff08\u30e9\u30f3\u30c0\u30e0\u30b5\u30d0\u30a4\u30d0\u30eb\u30d5\u30a9\u30ec\u30b9\u30c8 \/ RSF\uff09<\/strong> \u3092\u7528\u3044\u308c\u3070\u3001PH\u4eee\u5b9a\u306e\u7834\u7dbb\u3084\u8907\u96d1\u306a\u975e\u7dda\u5f62\u95a2\u4fc2\u3092\u30af\u30ea\u30a2\u3057\u305f\u4e0a\u3067\u3001<strong>\u5909\u6570\u91cd\u8981\u5ea6\uff08VIMP\uff09<\/strong>\u3001<strong>Minimal Depth\uff08\u6700\u5c0f\u6df1\u3055\uff09<\/strong>\u3001<strong>Partial Dependence Plot\uff08\u90e8\u5206\u4f9d\u5b58\u30d7\u30ed\u30c3\u30c8\uff09<\/strong> \u3068\u3044\u30463\u3064\u306e\u5f37\u529b\u306a\u6b66\u5668\u306b\u3088\u3063\u3066\u3001Cox\u56de\u5e30\u4ee5\u4e0a\u306b\u67d4\u8edf\u304b\u3064\u8aac\u5f97\u529b\u3092\u6301\u3063\u3066\u5371\u967a\u56e0\u5b50\u3092\u8a3c\u660e\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u672c\u8a18\u4e8b\u3067\u306f\u3001RSF\u306e\u76f4\u611f\u7684\u306a\u4ed5\u7d44\u307f\u304b\u3089\u3001\u6a5f\u68b0\u5b66\u7fd2\u306e\u521d\u5b66\u8005\u306b\u5411\u3051\u305f\u57fa\u672c\u7528\u8a9e\u89e3\u8aac\u3001\u6a19\u6e96\u30d1\u30c3\u30b1\u30fc\u30b8 <code>randomForestSRC<\/code> \u306b\u3088\u308b\u5b9f\u8df5\u624b\u9806\u3001\u305d\u3057\u3066\u8ad6\u6587\u3067\u305d\u306e\u307e\u307e\u4f7f\u3048\u308b\u82f1\u6587\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u307e\u3067\u3092\u5fb9\u5e95\u7684\u306b\u89e3\u8aac\u3059\u308b\u3002<\/p>\n\n\n\n<!--more-->\n\n\n\n<h2 class=\"wp-block-heading\">1. \u306a\u305c\u591a\u5909\u91cfCox\u56de\u5e30\u306f\u9650\u754c\u3092\u8fce\u3048\u308b\u306e\u304b\uff1f\uff08PH\u4eee\u5b9a\u7834\u7dbb\u3068\u9ad8\u6b21\u5143\u30c7\u30fc\u30bf\u306e\u58c1\uff09<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u533b\u5b66\u7814\u7a76\u306b\u304a\u3044\u3066\u6700\u3082\u5e83\u304f\u4f7f\u308f\u308c\u3066\u3044\u308b\u591a\u5909\u91cf\u89e3\u6790\u624b\u6cd5\u306fCox\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u30e2\u30c7\u30eb\u3067\u3042\u308b\u3002\u3057\u304b\u3057\u3001\u60a3\u8005\u80cc\u666f\u306e\u8907\u96d1\u5316\u306b\u4f34\u3044\u3001\u5f93\u6765\u306eCox\u56de\u5e30\u3067\u306f\u5bfe\u5fdc\u3057\u304d\u308c\u306a\u3044\u300c3\u3064\u306e\u81f4\u547d\u7684\u306a\u9650\u754c\u300d\u304c\u9855\u5728\u5316\u3057\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">1. \u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\uff08PH\u4eee\u5b9a\uff09\u306e\u7834\u7dbb<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Cox\u56de\u5e30\u306f\u300c\u89b3\u5bdf\u671f\u9593\u3092\u901a\u3058\u3066\u30012\u7fa4\u9593\u306e\u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u5e38\u306b\u4e00\u5b9a\u3067\u3042\u308b\u300d\u3068\u3044\u3046\u6975\u3081\u3066\u53b3\u683c\u306a\u524d\u63d0\u6761\u4ef6\uff08\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\uff09\u306b\u57fa\u3065\u3044\u3066\u3044\u308b\u3002\u3057\u304b\u3057\u3001\u8ffd\u8de1\u671f\u9593\u304c\u9577\u304f\u306a\u3063\u305f\u308a\u3001\u6cbb\u7642\u85ac\u306e\u521d\u671f\u52b9\u679c\u3068\u9577\u671f\u52b9\u679c\u304c\u7570\u306a\u3063\u305f\u308a\u3059\u308b\u5834\u5408\u3001\u30cf\u30b6\u30fc\u30c9\u6bd4\u306f\u6642\u9593\u3068\u3068\u3082\u306b\u5909\u5316\u3059\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5171\u5909\u91cf\u304c20\u301c30\u9805\u76ee\u3092\u8d85\u3048\u308b\u3068\u3001\u3069\u308c\u304b\u8907\u6570\u306e\u5909\u6570\u3067PH\u4eee\u5b9a\u304c\u7834\u7dbb\u3057\u3001\u59a5\u5f53\u306a\u5358\u4e00\u306eHR\u3092\u7b97\u51fa\u3067\u304d\u306a\u304f\u306a\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u203b\u88dc\u8db3\uff08\u95a2\u9023\u8a18\u4e8b\uff09<\/strong>:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\u306e\u7834\u7dbb\u6642\u306e\u5bfe\u7b56\u3084RMST\u89e3\u6790\u306b\u3064\u3044\u3066\u306f <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/handling-non-proportional-hazards-in-cox-regression-with-r-a-guide-to-rmst-stratification-and-time-varying-models\/\">Cox\u56de\u5e30\u3067\u306e\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\u7834\u7dbb\u3068RMST<\/a> \u3092\u53c2\u7167\u3055\u308c\u305f\u3044\u3002<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">2. \u9ad8\u6b21\u4ea4\u4e92\u4f5c\u7528\u3068\u975e\u7dda\u5f62\u6027\u306e\u7121\u8996<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u81e8\u5e8a\u73fe\u5834\u3067\u306f\u300c\u5358\u72ec\u306e\u30de\u30fc\u30ab\u30fc\u4e0a\u6607\u300d\u3067\u306f\u306a\u304f\u3001\u300c70\u6b73\u4ee5\u4e0a\u306e\u9ad8\u8840\u5727\u60a3\u8005\u3067\u306e\u307f\u3001\u7279\u5b9a\u306e\u708e\u75c7\u30de\u30fc\u30ab\u30fc\u304c\u9ad8\u304f\u306a\u308b\u3068\u6b7b\u4ea1\u30ea\u30b9\u30af\u304c\u8df3\u306d\u4e0a\u304c\u308b\u300d\u3068\u3044\u3063\u305f\u6761\u4ef6\u4ed8\u304d\u306e\u30ea\u30b9\u30af\u4e0a\u6607\uff08\u9ad8\u6b21\u4ea4\u4e92\u4f5c\u7528\uff09\u304c\u983b\u7e41\u306b\u8d77\u3053\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5f93\u6765\u306eCox\u56de\u5e30\u3067\u3053\u308c\u3092\u8a55\u4fa1\u3059\u308b\u306b\u306f\u3001\u7814\u7a76\u8005\u304c\u624b\u52d5\u3067\u300c\u4ea4\u4e92\u4f5c\u7528\u9805\uff08Interaction terms\uff09\u300d\u3092\u30e2\u30c7\u30eb\u306b\u7d44\u307f\u8fbc\u3080\u5fc5\u8981\u304c\u3042\u308b\u304c\u3001\u3042\u3089\u304b\u3058\u3081\u5168\u3066\u306e\u7d44\u307f\u5408\u308f\u305b\u3092\u60f3\u5b9a\u3059\u308b\u3053\u3068\u306f\u4e8b\u5b9f\u4e0a\u4e0d\u53ef\u80fd\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">3. \u591a\u91cd\u5171\u7dda\u6027\u3068\u904e\u5b66\u7fd2\uff08Overfitting\uff09\u306e\u7f60<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u5171\u5909\u91cf\u304c\u80cc\u666f\u9805\u76ee\u3084\u691c\u67fb\u5024\u306a\u3069\u6570\u5341\u9805\u76ee\u306b\u53ca\u3076\u5834\u5408\u3001\u5909\u6570\u9593\u306e\u76f8\u95a2\uff08\u591a\u91cd\u5171\u7dda\u6027\uff09\u306b\u3088\u308a\u30e2\u30c7\u30eb\u304c\u4e0d\u5b89\u5b9a\u306b\u306a\u308b\u3002\u30b9\u30c6\u30c3\u30d7\u30ef\u30a4\u30ba\u6cd5\u306a\u3069\u3067\u7121\u7406\u3084\u308a\u5909\u6570\u9078\u629e\u3092\u884c\u3046\u3068\u3001\u30b5\u30f3\u30d7\u30eb\u30b5\u30a4\u30ba\u306b\u5bfe\u3057\u3066\u30e2\u30c7\u30eb\u304c\u8907\u96d1\u306b\u306a\u308a\u3059\u304e\u3001\u672a\u77e5\u306e\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u4e88\u6e2c\u80fd\u3092\u5931\u3046\u300c\u904e\u5b66\u7fd2\uff08Overfitting\uff09\u300d\u3092\u5f15\u304d\u8d77\u3053\u3059\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<h3 class=\"wp-block-heading\">\u3010\u30b3\u30e9\u30e01\u3011\u81e8\u5e8a\u533b\u306e\u305f\u3081\u306e\u6a5f\u68b0\u5b66\u7fd2\u7528\u8a9e\u96c6\uff08\u57fa\u672c\u7de8\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u6a5f\u68b0\u5b66\u7fd2\u306b\u99b4\u67d3\u307f\u306e\u306a\u3044\u8aad\u8005\u5411\u3051\u306b\u3001\u672c\u8a18\u4e8b\u3067\u767b\u5834\u3059\u308b\u57fa\u672c\u6982\u5ff5\u3092\u81e8\u5e8a\u7684\u306a\u30a4\u30e1\u30fc\u30b8\u3067\u89e3\u8aac\u3059\u308b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u6c7a\u5b9a\u6728\uff08Decision Tree\uff09<\/strong>:\u300c\u5e74\u9f62 &gt; 65 \u304b\uff1f Yes\/No\u300d$\\rightarrow$\u300cCRP &gt; 2.0 \u304b\uff1f Yes\/No\u300d\u306e\u3088\u3046\u306b\u3001YES\/NO \u306e\u8cea\u554f\uff08\u5206\u5c90\uff09\u3092\u7e70\u308a\u8fd4\u3057\u3066\u60a3\u8005\u3092\u30ea\u30b9\u30af\u3054\u3068\u306b\u30b0\u30eb\u30fc\u30d7\u5206\u3051\u3059\u308b\u30d5\u30ed\u30fc\u30c1\u30e3\u30fc\u30c8\u3002<\/li>\n\n\n\n<li><strong>\u30a2\u30f3\u30b5\u30f3\u30d6\u30eb\u5b66\u7fd2\uff08Ensemble Learning\uff09<\/strong>:1\u4eba\u306e\u533b\u5e2b\uff081\u672c\u306e\u6c7a\u5b9a\u6728\uff09\u306e\u8a3a\u65ad\u306b\u983c\u308b\u306e\u3067\u306f\u306a\u304f\u30011000\u4eba\u306e\u533b\u5e2b\uff081000\u672c\u306e\u6c7a\u5b9a\u6728\uff09\u306b\u5c11\u3057\u305a\u3064\u7570\u306a\u308b\u60a3\u8005\u30c7\u30fc\u30bf\u3092\u898b\u305b\u3066\u5224\u5b9a\u3055\u305b\u3001\u305d\u306e<strong>\u300c\u591a\u6570\u6c7a\uff08\u5e73\u5747\uff09\u300d<\/strong>\u3067\u6700\u7d42\u7684\u306a\u30ea\u30b9\u30af\u4e88\u6e2c\u3092\u5f3e\u304d\u51fa\u3059\u624b\u6cd5\u3002\u300c\u68ee\uff08Forest\uff09\u300d\u3068\u547c\u3076\u306e\u306f\u3001\u7121\u6570\u306e\u300c\u6728\uff08Tree\uff09\u300d\u304c\u96c6\u307e\u3063\u3066\u3044\u308b\u305f\u3081\u3067\u3042\u308b\u3002<\/li>\n\n\n\n<li><strong>\u30d6\u30fc\u30c8\u30b9\u30c8\u30e9\u30c3\u30d7\u6a19\u672c\uff08Bootstrap Sample\uff09<\/strong>: \u5143\u306e\u81e8\u5e8a\u30c7\u30fc\u30bf\u304b\u3089\u91cd\u8907\u3092\u8a31\u3057\u3066\u30e9\u30f3\u30c0\u30e0\u306b\u60a3\u8005\u3092\u62bd\u51fa\uff08\u5fa9\u5143\u62bd\u51fa\uff09\u3057\u3001\u5c11\u3057\u305a\u3064\u9854\u3076\u308c\u306e\u9055\u3046\u300c\u5b66\u7fd2\u7528\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u300d\u3092\u4f55\u30d1\u30bf\u30fc\u30f3\u3082\u4f5c\u6210\u3059\u308b\u624b\u6cd5\u3002<\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n<h2 class=\"wp-block-heading\">2. Random Survival Forest (RSF) \u306e\u4ed5\u7d44\u307f\uff1a\u6253\u5207\u308a\u30c7\u30fc\u30bf\u3092\u6c7a\u5b9a\u6728\u3067\u6271\u3046\u6280\u6cd5<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u901a\u5e38\u306eRandom Forest\u306f\u300c\u304c\u3093\u767a\u751f\uff081\uff09\/ \u306a\u3057\uff080\uff09\u300d\u306e\u3088\u3046\u306a\u4e8c\u5024\u5206\u985e\u3084\u9023\u7d9a\u5024\u306e\u4e88\u6e2c\u3092\u884c\u3046\u624b\u6cd5\u3067\u3042\u308a\u3001\u8ffd\u8de1\u671f\u9593\uff08Time\uff09\u3084\u300c\u6253\u5207\u308a\uff08Censoring\uff09\u300d\u304c\u5b58\u5728\u3059\u308b\u751f\u5b58\u6642\u9593\u89e3\u6790\u306b\u306f\u305d\u306e\u307e\u307e\u9069\u7528\u3067\u304d\u306a\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u306e\u554f\u984c\u3092\u89e3\u6c7a\u3057\u305f\u306e\u304c <strong>Random Survival Forest\uff08RSF\uff09<\/strong> \u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u6253\u5207\u308a\u3092\u6b63\u3057\u304f\u6271\u3046\u300cLog-rank split rule\uff08\u30ed\u30b0\u30e9\u30f3\u30af\u5206\u5272\u30eb\u30fc\u30eb\uff09\u300d<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">RSF\u304c\u751f\u5b58\u6642\u9593\u30c7\u30fc\u30bf\u3092\u6b63\u3057\u304f\u6271\u3048\u308b\u79d8\u5bc6\u306f\u3001\u6c7a\u5b9a\u6728\u3092\u4f38\u3070\u3057\u3066\u3044\u304f\u969b\u306e\u300c\u5206\u5c90\u30eb\u30fc\u30eb\u300d\u306b\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e00\u822c\u7684\u306a\u6c7a\u5b9a\u6728\u304c\u30c7\u30fc\u30bf\u306e\u4e0d\u7d14\u5ea6\uff08Gini\u4fc2\u6570\u306a\u3069\uff09\u3092\u4e0b\u3052\u308b\u3088\u3046\u306b\u5206\u3051\u308b\u306e\u306b\u5bfe\u3057\u3001RSF\u3067\u306f\u5404\u30ce\u30fc\u30c9\uff08\u5206\u5c90\u70b9\uff09\u3067 <strong>Log-rank split rule\uff08\u30ed\u30b0\u30e9\u30f3\u30af\u5206\u5272\u30eb\u30fc\u30eb\uff09<\/strong> \u3092\u63a1\u7528\u3059\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3064\u307e\u308a\u3001\u300c\u60a3\u8005\u7fa4\u3092\u3069\u306e\u3088\u3046\u306a\u6761\u4ef6\uff08\u4f8b: \u8840\u5727 140 \u4ee5\u4e0a vs \u672a\u6e80\uff09\u30672\u5206\u5272\u3057\u305f\u3068\u304d\u306b\u30012\u7fa4\u9593\u306e\u30ed\u30b0\u30e9\u30f3\u30af\u7d71\u8a08\u91cf\u304c\u6700\u3082\u5927\u304d\u304f\u306a\u308b\u304b\uff08\u751f\u5b58\u66f2\u7dda\u304c\u6700\u3082\u5927\u304d\u304f\u96e2\u308c\u308b\u304b\uff09\u300d\u3092\u691c\u7d22\u3057\u3001\u5206\u5c90\u70b9\u3092\u4f5c\u6210\u3059\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u6253\u5207\u308a\u30c7\u30fc\u30bf\u3092\u6b63\u3057\u304f\u8003\u616e\u3057\u305f\u30ed\u30b0\u30e9\u30f3\u30af\u691c\u5b9a\u306b\u57fa\u3065\u3044\u3066\u6a39\u5f62\u56f3\uff08Survival Tree\uff09\u3092\u69cb\u7bc9\u3059\u308b\u305f\u3081\u3001\u6253\u5207\u308a\u3092\u542b\u3080\u81e8\u5e8a\u30c7\u30fc\u30bf\u304b\u3089\u3067\u3082\u6b63\u78ba\u306a\u7d2f\u7a4d\u30cf\u30b6\u30fc\u30c9\u3084\u4e88\u6e2c\u751f\u5b58\u7387\u3092\u30a2\u30f3\u30b5\u30f3\u30d6\u30eb\uff08\u96c6\u7d04\uff09\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<div id=\"biost-720854257\" class=\"biost- biost-entity-placement\"><p style=\"text-align: center;\"><span style=\"font-size: 20px;\"><strong><a href=\"https:\/\/best-biostatistics.com\/kmhl\">\uff1e\uff1e\u3082\u3046\u7d71\u8a08\u3067\u60a9\u3080\u306e\u306f\u7d42\u308f\u308a\u306b\u3057\u307e\u305b\u3093\u304b\uff1f\u00a0<\/a><\/strong><\/span><\/p>\r\n<a href=\"https:\/\/best-biostatistics.com\/kmhl\"><img class=\"aligncenter wp-image-2794 size-full\" src=\"https:\/\/best-biostatistics.com\/wp\/wp-content\/uploads\/2023\/11\/bn_r_03.png\" alt=\"\" width=\"500\" height=\"327\" \/><\/a>\r\n<p style=\"text-align: center;\"><span style=\"color: #ff0000; font-size: 20px;\"><strong><span class=\"marker2\">\u21911\u4e07\u4eba\u4ee5\u4e0a\u306e\u533b\u7642\u5f93\u4e8b\u8005\u304c\u8cfc\u8aad\u4e2d<\/span><\/strong><\/span><\/p><\/div><h2 class=\"wp-block-heading\">3. \u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u51fa\u306a\u304f\u3066\u3082\u67fb\u8aad\u8005\u3092\u9ed9\u3089\u305b\u308b\u300c3\u3064\u306e\u6b66\u5668\u300d<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u300cRSF\u306f\u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u51fa\u529b\u3055\u308c\u306a\u3044\u304b\u3089\u3001\u533b\u5b66\u8ad6\u6587\uff08\u67fb\u8aad\uff09\u3067\u4f7f\u3048\u306a\u3044\u306e\u3067\u306f\u306a\u3044\u304b\uff1f\u300d\u3068\u3044\u3046\u8cea\u554f\u3092\u3088\u304f\u53d7\u3051\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3057\u304b\u3057\u3001\u3053\u308c\u306f\u5b8c\u5168\u306a\u8aa4\u89e3\u3067\u3042\u308b\u3002\u67fb\u8aad\u8005\u304c\u672c\u8cea\u7684\u306b\u77e5\u308a\u305f\u3044\u306e\u306f\u300c\u5358\u4e00\u306eHR\u3068\u3044\u3046\u6570\u5024\u300d\u305d\u306e\u3082\u306e\u3067\u306f\u306a\u304f\u3001<strong>\u300c\u3069\u306e\u56e0\u5b50\u304c\u771f\u306e\u4e88\u5f8c\u4e88\u6e2c\u56e0\u5b50\u304b\u300d\u300c\u305d\u306e\u56e0\u5b50\u304c\u30ea\u30b9\u30af\u306b\u3069\u3046\u5f71\u97ff\u3057\u3066\u3044\u308b\u304b\u300d\u300c\u30e2\u30c7\u30eb\u306e\u4e88\u6e2c\u7cbe\u5ea6\u306f\u9ad8\u3044\u304b\u300d<\/strong> \u306e3\u70b9\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">RSF\u3067\u306f\u3001Cox\u56de\u5e30\u306eHR\u306b\u4ee3\u308f\u308b\u300c3\u3064\u306e\u5f37\u529b\u306a\u53ef\u8996\u5316\u30c4\u30fc\u30eb\u300d\u3092\u7528\u3044\u308b\u3053\u3068\u3067\u3001\u67fb\u8aad\u8005\u3092\u5341\u5206\u306b\u7d0d\u5f97\u3055\u305b\u308b\u30ed\u30b8\u30c3\u30af\u3092\u69cb\u7bc9\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u6b66\u56681\uff1aVIMP\uff08Variable Importance \/ \u5909\u6570\u91cd\u8981\u5ea6\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u7279\u5b9a\u306e\u5909\u6570\uff08\u4f8b: Age\uff09\u306e\u5024\u3092\u30e9\u30f3\u30c0\u30e0\u306b\u30b7\u30e3\u30c3\u30d5\u30eb\uff08\u7834\u68c4\uff09\u3057\u305f\u3068\u304d\u306b\u3001\u30e2\u30c7\u30eb\u5168\u4f53\u306e\u4e88\u6e2c\u7cbe\u5ea6\u304c\u3069\u308c\u304f\u3089\u3044\u60aa\u5316\u3059\u308b\u304b\u3092\u6e2c\u5b9a\u3059\u308b\u6307\u6a19\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3082\u3057\u3001\u3042\u308b\u5909\u6570\u3092\u7834\u68c4\u3057\u305f\u3068\u304d\u306b\u4e88\u6e2c\u7cbe\u5ea6\u304c\u8457\u3057\u304f\u4f4e\u4e0b\uff08\u30a8\u30e9\u30fc\u7387\u304c\u4e0a\u6607\uff09\u3057\u305f\u5834\u5408\u3001\u305d\u306e\u5909\u6570\u306f\u300c\u4e88\u5f8c\u4e88\u6e2c\u306b\u6975\u3081\u3066\u91cd\u8981\u306a\u56e0\u5b50\u3067\u3042\u308b\u300d\u3068\u5224\u5b9a\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u6b66\u56682\uff1aMinimal Depth\uff08\u6700\u5c0f\u6df1\u3055\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u6c7a\u5b9a\u6728\u306e\u4e2d\u3067\u3001\u305d\u306e\u5909\u6570\u304c\u300c\u3069\u308c\u304f\u3089\u3044\u6d45\u3044\uff08\u521d\u671f\u306e\uff09\u968e\u5c64\u300d\u3067\u5206\u5272\u306b\u4f7f\u308f\u308c\u305f\u304b\u3092\u793a\u3059\u6307\u6a19\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e88\u5f8c\u3092\u6c7a\u5b9a\u3065\u3051\u308b\u5f37\u56fa\u306a\u30ea\u30b9\u30af\u56e0\u5b50\u307b\u3069\u3001\u6c7a\u5b9a\u6728\u306e\u4e0a\u90e8\uff08\u6d45\u3044\u968e\u5c64\uff09\u3067\u983b\u7e41\u306b\u9078\u629e\u3055\u308c\u308b\u3002VIMP\u3068Minimal Depth\u306e\u53cc\u65b9\u3067\u4e0a\u4f4d\u306b\u5165\u3063\u305f\u56e0\u5b50\u306f\u3001\u6975\u3081\u3066\u78ba\u5b9f\u6027\u306e\u9ad8\u3044\u30ea\u30b9\u30af\u56e0\u5b50\u3068\u3057\u3066\u4e3b\u5f35\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u6b66\u56683\uff1aPartial Dependence Plot\uff08PDP \/ \u90e8\u5206\u4f9d\u5b58\u30d7\u30ed\u30c3\u30c8\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ed6\u306e\u3059\u3079\u3066\u306e\u5909\u6570\u306e\u5f71\u97ff\u3092\u4e00\u5b9a\u306b\u4fdd\u3063\u305f\uff08\u5468\u8fba\u5316\u3057\u305f\uff09\u4e0a\u3067\u3001\u7279\u5b9a\u306e\u30ea\u30b9\u30af\u56e0\u5b50\uff08\u4f8b: D-dimer\u5024\u3001CRP\u5024\u3001\u5e74\u9f62\uff09\u306e\u5909\u5316\u306b\u4f34\u3044\u3001\u30a2\u30f3\u30b5\u30f3\u30d6\u30eb\u6b7b\u4ea1\u30ea\u30b9\u30af\u3084\u4e88\u6e2c\u751f\u5b58\u7387\u304c\u3069\u3046\u63a8\u79fb\u3059\u308b\u304b\u3092\u63cf\u753b\u3059\u308b\u30d7\u30ed\u30c3\u30c8\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u300c\u691c\u67fb\u5024\u304c\u3007\u3007\u3092\u8d85\u3048\u308b\u3068\u6025\u6fc0\u306b\u6b7b\u4ea1\u30ea\u30b9\u30af\u304c\u8df3\u306d\u4e0a\u304c\u308b\u300d\u3068\u3044\u3063\u305f\u975e\u7dda\u5f62\u306a\u95be\u5024\u52b9\u679c\uff08Threshold effect\uff09\u3092\u8996\u899a\u7684\u306b\u63d0\u793a\u3067\u304d\u308b\u305f\u3081\u3001\u5358\u4e00\u306eHR\u5024\u3088\u308a\u3082\u81e8\u5e8a\u7684\u306b\u4fa1\u5024\u306e\u9ad8\u3044\u60c5\u5831\u3092\u63d0\u4f9b\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<h3 class=\"wp-block-heading\">\u3010\u30b3\u30e9\u30e02\u3011\u81e8\u5e8a\u533b\u306e\u305f\u3081\u306e\u6a5f\u68b0\u5b66\u7fd2\u7528\u8a9e\u96c6\uff08\u8a55\u4fa1\u30fb\u89e3\u91c8\u7de8\uff09<\/h3>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Out-of-Bag (OOB) \u30c7\u30fc\u30bf<\/strong>: \u30d6\u30fc\u30c8\u30b9\u30c8\u30e9\u30c3\u30d7\u6a19\u672c\u3092\u4f5c\u6210\u3059\u308b\u969b\u3001<strong>\u300c\u305f\u307e\u305f\u307e\u9078\u3070\u308c\u306a\u304b\u3063\u305f\u7d0437%\u306e\u60a3\u8005\u30c7\u30fc\u30bf\u300d<\/strong> \u306e\u3053\u3068\u3002<strong>\u3010\u306a\u305c\u7d0437%\u306b\u306a\u308b\u306e\u304b\uff1f\u3011<\/strong> \u30b5\u30f3\u30d7\u30eb\u30b5\u30a4\u30ba$N$\u4eba\u306e\u30c7\u30fc\u30bf\u304b\u3089\u300c\u91cd\u8907\u3092\u8a31\u3057\u3066$N$\u56de\u30e9\u30f3\u30c0\u30e0\u306b\u9078\u3076\uff08\u5fa9\u5143\u62bd\u51fa\uff09\u300d\u3068\u304d\u3001\u3042\u308b\u60a3\u8005\u304c1\u56de\u306e\u62bd\u51fa\u3067\u300e\u9078\u3070\u308c\u306a\u3044\u78ba\u7387\u300f\u306f$(1 &#8211; 1\/N)$\u3067\u3042\u308b\u3002\u3053\u308c\u3092$N$\u56de\u7e70\u308a\u8fd4\u3057\u305f\u3068\u304d\u306b\u300e\u4e00\u5ea6\u3082\u9078\u3070\u308c\u306a\u3044\u78ba\u7387\u300f\u306f$(1 &#8211; 1\/N)^N$\u3068\u306a\u308b\u3002 \u30b5\u30f3\u30d7\u30eb\u30b5\u30a4\u30ba$N$\u304c\u5341\u5206\u5927\u304d\u3044\u81e8\u5e8a\u30c7\u30fc\u30bf\u3067\u306f\u3001\u6570\u5b66\u306e\u6027\u8cea\u4e0a\u3001\u3053\u306e\u5024\u306f\u30cd\u30a4\u30d4\u30a2\u6570\u306e\u9006\u6570$1\/e \\approx 0.368$\uff08<strong>\u7d0436.8%<\/strong>$\\approx$<strong>37%<\/strong>\uff09\u3078\u53ce\u675f\u3059\u308b\u3002RSF\u3067\u306f\u3001\u3053\u306e\u7d0437%\u306e\u672a\u5b66\u7fd2\u30c7\u30fc\u30bf\uff08OOB\uff09\u3092\u30c6\u30b9\u30c8\u7528\u30c7\u30fc\u30bf\u3068\u3057\u3066\u7528\u3044\u3066\u7cbe\u5ea6\u691c\u8a3c\u3092\u884c\u3046\u305f\u3081\u3001\u5225\u9014\u691c\u8a3c\u7528\u30c7\u30fc\u30bf\u3092\u7528\u610f\u3057\u306a\u304f\u3066\u3082\u904e\u5b66\u7fd2\uff08Overfitting\uff09\u306e\u306a\u3044\u5ba2\u89b3\u7684\u306a\u4e88\u6e2c\u7cbe\u5ea6\u3092\u7b97\u51fa\u3067\u304d\u308b\u3002<\/li>\n\n\n\n<li><strong>C-index (Harrell&#8217;s Concordance Index)<\/strong>:\u300c\u3088\u308a\u9ad8\u30ea\u30b9\u30af\u3068\u4e88\u6e2c\u3055\u308c\u305f\u60a3\u8005\u304c\u3001\u5b9f\u969b\u306b\u65e9\u304f\u30a4\u30d9\u30f3\u30c8\u3092\u8d77\u3053\u3057\u305f\u304b\u300d\u3068\u3044\u3046\u4e00\u81f4\u7387\u3092\u793a\u3059\u6307\u6a19\uff080.5\u301c1.0\uff09\u30020.5\u3067\u30b3\u30a4\u30f3\u6295\u3052\u3068\u540c\u7b49\u30010.7\u4ee5\u4e0a\u3067\u826f\u597d\u306a\u4e88\u6e2c\u80fd\u30010.8\u4ee5\u4e0a\u3067\u6975\u3081\u3066\u9ad8\u7cbe\u5ea6\u3068\u5224\u5b9a\u3055\u308c\u308b\u3002<\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n<h2 class=\"wp-block-heading\">4. \u3010\u5b9f\u8df5R\u30b3\u30fc\u30c9\u3011<code>randomForestSRC<\/code> \u306b\u3088\u308b RSF\u30e2\u30c7\u30eb\u69cb\u7bc9\u3068\u53ef\u8996\u5316\uff08Step by Step\uff09<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u3053\u304b\u3089\u306f\u3001CRAN\u3067\u6700\u3082\u5b89\u5b9a\u3057\u3066\u30e1\u30f3\u30c6\u3055\u308c\u3066\u3044\u308b\u6a19\u6e96\u30d1\u30c3\u30b1\u30fc\u30b8 <code>randomForestSRC<\/code> \u3092\u7528\u3044\u3066\u3001RSF\u30e2\u30c7\u30eb\u69cb\u7bc9\u304b\u3089\u7cbe\u5ea6\u8a55\u4fa1\u3001\u8981\u56e0\u62bd\u51fa\u307e\u3067\u3092 Step by Step \u3067\u89e3\u8aac\u3059\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 1: \u7591\u4f3c\u81e8\u5e8a\u30c7\u30fc\u30bf\u306e\u4f5c\u6210\u3068\u78ba\u8a8d<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u60a3\u8005200\u4eba\u5206\u3001\u8ffd\u8de1\u6642\u9593\uff08<code>time<\/code>\uff09\u3001\u30a4\u30d9\u30f3\u30c8\uff08<code>status<\/code>\uff09\u3001\u80cc\u666f\u56e0\u5b5010\u9805\u76ee\u306e\u7591\u4f3c\u81e8\u5e8a\u30c7\u30fc\u30bf\u3092\u751f\u6210\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>library(randomForestSRC)\nlibrary(tidyverse)\n\n# \u4e71\u6570\u30b7\u30fc\u30c9\u306e\u8a2d\u5b9a\nset.seed(123)\nn &lt;- 200\n\n# \u7591\u4f3c\u81e8\u5e8a\u30c7\u30fc\u30bf\u306e\u4f5c\u6210\ndf_rsf &lt;- tibble(\n  id = 1:n,\n  time = round(runif(n, min = 1, max = 60), 1),\n  status = sample(c(0, 1), size = n, replace = TRUE, prob = c(0.4, 0.6)),\n  age = round(rnorm(n, mean = 65, sd = 10)),\n  gender = factor(sample(c(\"Male\", \"Female\"), n, replace = TRUE)),\n  bmi = round(rnorm(n, mean = 24, sd = 4), 1),\n  sbp = round(rnorm(n, mean = 135, sd = 15)),\n  crp = round(rgamma(n, shape = 2, scale = 1.5), 2),\n  eGFR = round(rnorm(n, mean = 70, sd = 15), 1),\n  bnp = round(rlnorm(n, meanlog = 4.5, sdlog = 0.8), 1),\n  statine = factor(sample(c(\"No\", \"Yes\"), n, replace = TRUE)),\n  smoking = factor(sample(c(\"Never\", \"Former\", \"Current\"), n, replace = TRUE))\n)\n\n# \u5148\u982d\u884c\u306e\u78ba\u8a8d\nhead(df_rsf)\n<\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">Step 2: RSF\u30e2\u30c7\u30eb\u306e\u69cb\u7bc9\u3068 OOB C-index \u306b\u3088\u308b\u7cbe\u5ea6\u8a55\u4fa1<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><code>rfsrc()<\/code> \u95a2\u6570\u3092\u7528\u3044\u3066 <code>ntree = 1000<\/code>\uff081000\u672c\u306e\u6c7a\u5b9a\u6728\uff09\u306e\u30d5\u30a9\u30ec\u30b9\u30c8\u3092\u69cb\u7bc9\u3057\u3001Out-of-Bag (OOB) \u30c7\u30fc\u30bf\u306b\u57fa\u3065\u304f C-index\uff08\u4e88\u6e2c\u7cbe\u5ea6\uff09\u3092\u51fa\u529b\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># RSF\u30e2\u30c7\u30eb\u306e\u69cb\u7bc9\nrsf_fit &lt;- rfsrc(\n  Surv(time, status) ~ .,\n  data = df_rsf %&gt;% select(-id),\n  ntree = 1000,\n  splitrule = \"logrank\",\n  importance = TRUE,\n  seed = 123\n)\n\n# \u30e2\u30c7\u30eb\u6982\u8981\u306e\u51fa\u529b\nprint(rsf_fit)\n\n# OOB C-index \u306e\u78ba\u8a8d (1 - OOB Error Rate)\noob_c_index &lt;- 1 - rsf_fit$err.rate&#91;rsf_fit$ntree]\ncat(sprintf(\"Out-of-Bag (OOB) C-index: %.3f\\n\", oob_c_index))\n<\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">Step 3: \u5909\u6570\u91cd\u8981\u5ea6\uff08VIMP\uff09\u3068 Minimal Depth \u306e\u62bd\u51fa\u30fb\u6bd4\u8f03<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><code>vimp()<\/code> \u304a\u3088\u3073 <code>max.subtree()<\/code> \u3092\u7528\u3044\u3066\u3001\u91cd\u8981\u5909\u6570\u3092\u62bd\u51fa\u30fb\u30e9\u30f3\u30ad\u30f3\u30b0\u5316\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># VIMP (Variable Importance) \u306e\u53d6\u5f97\nvimp_res &lt;- vimp(rsf_fit)$importance\n\n# Minimal Depth (\u6700\u5c0f\u6df1\u3055) \u306e\u62bd\u51fa\nmax_sub_res &lt;- max.subtree(rsf_fit)\nmin_depth_res &lt;- max_sub_res$order&#91;,1]\n\n# VIMP\u3068Minimal Depth\u306e\u7d50\u679c\u3092\u7d71\u5408\ndf_importance &lt;- tibble(\n  variable = names(vimp_res),\n  vimp = as.numeric(vimp_res)\n) %>%\n  left_join(\n    tibble(\n      variable = names(min_depth_res),\n      min_depth = as.numeric(min_depth_res)\n    ),\n    by = \"variable\"\n  ) %>%\n  arrange(desc(vimp))\n\nprint(df_importance)\n<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u3010\u7d50\u679c\u306e\u8aad\u307f\u65b9\u3011<\/strong>VIMP\u306e\u5024\u304c\u5927\u304d\u3044\u307b\u3069\u4e88\u6e2c\u80fd\u3078\u306e\u8ca2\u732e\u5ea6\u304c\u9ad8\u304f\u3001Minimal Depth\u306e\u5024\u304c\u5c0f\u3055\u3044\uff08\uff10\u306b\u8fd1\u3044\uff09\u307b\u3069\u6c7a\u5b9a\u6728\u306e\u6839\u5143\u8fd1\u304f\u3067\u65e9\u671f\u306b\u5206\u985e\u306b\u4f7f\u308f\u308c\u305f\u91cd\u8981\u306a\u5909\u6570\u3068\u306a\u308b\u3002\u53cc\u65b9\u306e\u6307\u6a19\u3067\u4e0a\u4f4d\u306b\u5165\u3063\u305f\u5909\u6570\u3053\u305d\u304c\u3001\u771f\u306b\u78ba\u5b9f\u6027\u306e\u9ad8\u3044\u4e88\u5f8c\u4e88\u6e2c\u56e0\u5b50\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 4: Partial Dependence Plot (PDP) \u306b\u3088\u308b\u975e\u7dda\u5f62\u30ea\u30b9\u30af\u63a8\u79fb\u306e\u63cf\u753b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e0a\u4f4d\u306b\u62bd\u51fa\u3055\u308c\u305f\u91cd\u8981\u306a\u9023\u7d9a\u5909\u6570\u306b\u3064\u3044\u3066\u3001\u6570\u5024\u306e\u5909\u5316\u306b\u4f34\u3046\u6b7b\u4ea1\u30ea\u30b9\u30af\u306e\u975e\u7dda\u5f62\u306a\u5909\u5316\u3092\u63cf\u753b\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># PDP\uff08\u90e8\u5206\u4f9d\u5b58\u30d7\u30ed\u30c3\u30c8\uff09\u306e\u63cf\u753b\nplot.variable(\n  rsf_fit,\n  xvar.names = c(\"eGFR\", \"sbp\", \"bnp\"),\n  partial = TRUE,\n  plots.per.page = 3\n)\n<\/code><\/pre>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" width=\"1024\" height=\"483\" src=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/08\/image-10.png\" alt=\"\" class=\"wp-image-4734\" srcset=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/08\/image-10.png 1024w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/08\/image-10-300x142.png 300w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/08\/image-10-768x362.png 768w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">5. \u8ad6\u6587\u3067\u305d\u306e\u307e\u307e\u4f7f\u3048\u308b\u82f1\u6587\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\uff08Methods &amp; Results\uff09<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u8ad6\u6587\u306e Methods \u304a\u3088\u3073 Results \u30bb\u30af\u30b7\u30e7\u30f3\u306b\u305d\u306e\u307e\u307e\u8a18\u8ff0\u3067\u304d\u308b\u6a19\u6e96\u7684\u306a\u82f1\u6587\u8868\u73fe\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8\u3092\u63d0\u793a\u3059\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Methods \u30bb\u30af\u30b7\u30e7\u30f3\u82f1\u6587\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">&#8220;To account for potential non-proportional hazards, high-dimensional covariates, and non-linear interactions without making parametric assumptions, a Random Survival Forest (RSF) model was constructed using the <code>randomForestSRC<\/code> package in R. The forest was built with 1,000 survival trees using the log-rank splitting rule. Model performance was evaluated using the Out-of-Bag (OOB) Harrell\u2019s concordance index (C-index) to prevent overfitting. Variable importance (VIMP) and minimal depth analysis were utilized to identify and rank key prognostic factors. The non-linear relationship between continuous clinical variables and mortality risk was visualized using Partial Dependence Plots (PDP).&#8221;<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Results \u30bb\u30af\u30b7\u30e7\u30f3\u82f1\u6587\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">&#8220;The Random Survival Forest model demonstrated robust predictive accuracy, achieving an Out-of-Bag C-index of [0.545]. Among the evaluated clinical factors, [eGFR, SBP, and BNP,] were identified as the most dominant prognostic factors according to both Variable Importance (VIMP) and minimal depth metrics. Partial Dependence Plots revealed a non-linear threshold effect, demonstrating a sharp increase in predicted mortality risk when [eGFR levels dropped below 60 mL\/min\/1.73 $\\mathrm{m}^2$].&#8221;<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">6. \u307e\u3068\u3081\uff06\u4e00\u62ec\u30b3\u30d4\u30da\u7528R\u30b9\u30af\u30ea\u30d7\u30c8<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">\u672c\u8a18\u4e8b\u306e\u30dd\u30a4\u30f3\u30c8\u304a\u3055\u3089\u3044<\/h3>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>PH\u4eee\u5b9a\u7834\u7dbb\u3084\u591a\u9805\u76ee\u30c7\u30fc\u30bf\u306b\u306fRSF\u304c\u8d85\u5f37\u529b<\/strong>: Cox\u56de\u5e30\u306e\u9650\u754c\uff08\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\u306e\u7834\u7dbb\u3001\u975e\u7dda\u5f62\u30fb\u9ad8\u6b21\u4ea4\u4e92\u4f5c\u7528\u3001\u904e\u5b66\u7fd2\uff09\u3092\u81ea\u52d5\u3067\u56de\u907f\u3067\u304d\u308b\u3002<\/li>\n\n\n\n<li><strong>\u6253\u5207\u308a\u30c7\u30fc\u30bf\u3082Log-rank split\u3067\u5b8c\u74a7\u306b\u51e6\u7406<\/strong>: \u901a\u5e38\u306e\u6c7a\u5b9a\u6728\u3068\u7570\u306a\u308a\u3001\u30ed\u30b0\u30e9\u30f3\u30af\u5206\u5272\u30eb\u30fc\u30eb\u306b\u3088\u308a\u751f\u5b58\u6642\u9593\u30c7\u30fc\u30bf\u3092\u6b63\u78ba\u306b\u96c6\u7d04\u3059\u308b\u3002<\/li>\n\n\n\n<li><strong>HR\u304c\u51fa\u306a\u304f\u3066\u3082\u300c3\u3064\u306e\u6b66\u5668\u300d\u3067\u67fb\u8aad\u306b\u5bfe\u5fdc<\/strong>: VIMP\uff08\u91cd\u8981\u5ea6\uff09\u3001Minimal Depth\uff08\u6df1\u3055\uff09\u3001PDP\uff08\u90e8\u5206\u4f9d\u5b58\u30d7\u30ed\u30c3\u30c8\uff09\u3092\u63d0\u793a\u3059\u308b\u3053\u3068\u3067\u3001Cox\u56de\u5e30\u4ee5\u4e0a\u306b\u67d4\u8edf\u304b\u3064\u5b9f\u52d9\u7684\u306a\u30ea\u30b9\u30af\u8a55\u4fa1\u304c\u53ef\u80fd\u3002<\/li>\n<\/ol>\n\n\n\n<h3 class=\"wp-block-heading\">\u4e00\u62ec\u5b9f\u884c\u7528R\u30b9\u30af\u30ea\u30d7\u30c8<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ee5\u4e0b\u306e\u30b3\u30fc\u30c9\u3092\u30b3\u30d4\u30fc\uff06\u30da\u30fc\u30b9\u30c8\u3059\u308b\u3053\u3068\u3067\u3001\u7591\u4f3c\u30c7\u30fc\u30bf\u751f\u6210\u304b\u3089RSF\u30e2\u30c7\u30eb\u69cb\u7bc9\u3001OOB C-index\u7b97\u51fa\u3001VIMP\/Minimal Depth\u62bd\u51fa\u3001PDP\u63cf\u753b\u307e\u3067\u3092\u4e00\u6c17\u901a\u8cab\u3067\u8d70\u3089\u305b\u308b\u3053\u3068\u304c\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># ==============================================================================\n# Random Survival Forest (RSF) \u4e00\u62ec\u5b9f\u884c\u30b9\u30af\u30ea\u30d7\u30c8\n# \u30d1\u30c3\u30b1\u30fc\u30b8: randomForestSRC, tidyverse\n# ==============================================================================\n\nif (!requireNamespace(\"randomForestSRC\", quietly = TRUE)) install.packages(\"randomForestSRC\")\nif (!requireNamespace(\"tidyverse\", quietly = TRUE)) install.packages(\"tidyverse\")\n\nlibrary(randomForestSRC)\nlibrary(tidyverse)\n\n# 1. \u7591\u4f3c\u81e8\u5e8a\u30c7\u30fc\u30bf\u306e\u4f5c\u6210\nset.seed(123)\nn &lt;- 200\n\ndf_rsf &lt;- tibble(\n  id = 1:n,\n  time = round(runif(n, min = 1, max = 60), 1),\n  status = sample(c(0, 1), size = n, replace = TRUE, prob = c(0.4, 0.6)),\n  age = round(rnorm(n, mean = 65, sd = 10)),\n  gender = factor(sample(c(\"Male\", \"Female\"), n, replace = TRUE)),\n  bmi = round(rnorm(n, mean = 24, sd = 4), 1),\n  sbp = round(rnorm(n, mean = 135, sd = 15)),\n  crp = round(rgamma(n, shape = 2, scale = 1.5), 2),\n  eGFR = round(rnorm(n, mean = 70, sd = 15), 1),\n  bnp = round(rlnorm(n, meanlog = 4.5, sdlog = 0.8), 1),\n  statine = factor(sample(c(\"No\", \"Yes\"), n, replace = TRUE)),\n  smoking = factor(sample(c(\"Never\", \"Former\", \"Current\"), n, replace = TRUE))\n)\n\n# 2. RSF\u30e2\u30c7\u30eb\u306e\u69cb\u7bc9\ncat(\"\\n--- Random Survival Forest \u30e2\u30c7\u30eb\u69cb\u7bc9\u4e2d ---\\n\")\nrsf_fit &lt;- rfsrc(\n  Surv(time, status) ~ .,\n  data = df_rsf %>% select(-id),\n  ntree = 1000,\n  splitrule = \"logrank\",\n  importance = TRUE,\n  seed = 123\n)\n\n# 3. \u4e88\u6e2c\u7cbe\u5ea6\uff08OOB C-index\uff09\u306e\u8a55\u4fa1\noob_c_index &lt;- 1 - rsf_fit$err.rate&#91;rsf_fit$ntree]\ncat(sprintf(\"\\nOut-of-Bag (OOB) C-index: %.3f\\n\", oob_c_index))\n\n# 4. \u5909\u6570\u91cd\u8981\u5ea6 (VIMP) \u306e\u7b97\u51fa\nvimp_res &lt;- vimp(rsf_fit)$importance\n\n# Minimal Depth (\u6700\u5c0f\u6df1\u3055) \u306e\u62bd\u51fa\nmax_sub_res &lt;- max.subtree(rsf_fit)\nmin_depth_res &lt;- max_sub_res$order&#91;,1]\n\n# VIMP\u3068Minimal Depth\u306e\u7d50\u679c\u3092\u7d71\u5408\ndf_importance &lt;- tibble(\n  variable = names(vimp_res),\n  vimp = as.numeric(vimp_res)\n) %>%\n  left_join(\n    tibble(\n      variable = names(min_depth_res),\n      min_depth = as.numeric(min_depth_res)\n    ),\n    by = \"variable\"\n  ) %>%\n  arrange(desc(vimp))\n\nprint(df_importance)\n\ncat(\"\\n--- \u5909\u6570\u91cd\u8981\u5ea6 (VIMP) \u3068 Minimal Depth \u30e9\u30f3\u30ad\u30f3\u30b0 ---\\n\")\nprint(df_importance)\n\n# 5. PDP\uff08\u90e8\u5206\u4f9d\u5b58\u30d7\u30ed\u30c3\u30c8\uff09\u306e\u63cf\u753b\nplot.variable(\n  rsf_fit,\n  xvar.names = df_importance$variable&#91;1:3], # \u4e0a\u4f4d3\u5909\u6570\u3092\u30d7\u30ed\u30c3\u30c8\n  partial = TRUE,\n  plots.per.page = 3\n)\n<\/code><\/pre>\n\n\n\n<h2 class=\"wp-block-heading\">\u304a\u3059\u3059\u3081\u66f8\u7c4d<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/amzn.to\/4ga3oV8\">\u8ab0\u3082\u6559\u3048\u3066\u304f\u308c\u306a\u304b\u3063\u305f\u3000\u533b\u7642\u7d71\u8a08\u306e\u4f7f\u3044\u5206\u3051\u301c\u8ff7\u3044\u3084\u3059\u3044\u89e3\u6790\u624b\u6cd5\u306e\u9078\u3073\u65b9\u304c\uff0cR\u3067\u5b9f\u611f\u3057\u306a\u304c\u3089\u308f\u304b\u308b\uff01<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u300c\u5148\u751f\uff01\u591a\u5909\u91cfCox\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u30e2\u30c7\u30eb\u3092\u7d44\u3082\u3046\u3068\u3057\u305f\u3089\u3001\u6bd4\u4f8b\u30cf\u30b6\u30fc\u30c9\u6027\uff08PH\u4eee\u5b9a\uff09\u304c\u8907\u6570\u306e\u5909\u6570\u3067\u5d29\u308c\u3066\u3057\u307e\u3044\u3001\u5358\u4e00\u306e\u30cf\u30b6\u30fc\u30c9\u6bd4\uff08HR\uff09\u304c\u51fa\u305b\u306a\u304f\u306a\u308a\u307e\u3057\u305f\uff01\u304a\u307e\u3051\u306b\u80cc\u666f\u56e0\u5b50\u304c30\u9805\u76ee\u4ee5\u4e0a\u3042\u3063\u3066\u3001\u30b9\u30c6\u30c3\u30d7\u30ef\u30a4\u30ba\u6cd5\u3092\u5165\u308c\u3066\u3082 [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":4743,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"swell_btn_cv_data":"","_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[103],"tags":[],"class_list":["post-4731","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-103"],"jetpack_publicize_connections":[],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/08\/random-survival-forest-scaled.jpeg","_links":{"self":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4731","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/comments?post=4731"}],"version-history":[{"count":9,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4731\/revisions"}],"predecessor-version":[{"id":4742,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4731\/revisions\/4742"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/media\/4743"}],"wp:attachment":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/media?parent=4731"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/categories?post=4731"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/tags?post=4731"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}