{"id":8466,"date":"2026-05-19T16:28:43","date_gmt":"2026-05-19T07:28:43","guid":{"rendered":"https:\/\/since2020.jp\/media\/?p=8466"},"modified":"2026-05-19T16:28:43","modified_gmt":"2026-05-19T07:28:43","slug":"claude_improving_competition","status":"publish","type":"post","link":"https:\/\/since2020.jp\/media\/claude_improving_competition\/","title":{"rendered":"\u3010AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3011ClaudeCode\u3092\u4f7f\u7528\u3057\u3066\u30b3\u30f3\u30da\u306e\u7cbe\u5ea6\u6539\u5584\u3092\u81ea\u52d5\u5316\u3059\u308b\u3002"},"content":{"rendered":"\n

\u306f\u3058\u3081\u306b<\/h2>\n\n\n\n

\u8fd1\u5e74\u3001\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff08LLM\uff09\u3092\u6d3b\u7528\u3057\u305f\u81ea\u5f8b\u7684\u306a\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u6025\u901f\u306b\u666e\u53ca\u3057\u3066\u304a\u308a\u3001\u7279\u306b\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u958b\u767a\u306e\u9818\u57df\u306b\u304a\u3044\u3066\u30b3\u30fc\u30c9\u751f\u6210\u30fb\u4fee\u6b63\u30fb\u5b9f\u884c\u3092\u81ea\u52d5\u5316\u3059\u308b\u30c4\u30fc\u30eb\u3068\u3057\u3066\u5e83\u304f\u6d3b\u7528\u3055\u308c\u3066\u3044\u308b\u3002\u305d\u306e\u4ee3\u8868\u4f8b\u3068\u3057\u3066\u3001Anthropic\u304c\u958b\u767a\u3057\u305fClaude Code\u304c\u6319\u3052\u3089\u308c\u308b\u3002ClaudeCode\u306f\u30bf\u30fc\u30df\u30ca\u30eb\u4e0a\u3067\u52d5\u4f5c\u3059\u308b\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3067\u3042\u308a\u3001\u30d5\u30a1\u30a4\u30eb\u306e\u8aad\u307f\u66f8\u304d\u3084\u30b3\u30de\u30f3\u30c9\u306e\u5b9f\u884c\u3092\u542b\u3080\u8907\u96d1\u306a\u30bf\u30b9\u30af\u3092\u81ea\u5f8b\u7684\u306b\u51e6\u7406\u3059\u308b\u80fd\u529b\u3092\u6301\u3064\u3002<\/p>\n\n\n\n

\u307e\u305f\u3001\u30c7\u30fc\u30bf\u30b5\u30a4\u30a8\u30f3\u30b9\u30b3\u30f3\u30da\u30c6\u30a3\u30b7\u30e7\u30f3\u306f\u3001\u30c7\u30fc\u30bf\u306e\u63a2\u7d22\u30fb\u5206\u6790\u304b\u3089\u7279\u5fb4\u91cf\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\u30fb\u30e2\u30c7\u30eb\u69cb\u7bc9\u306b\u81f3\u308b\u307e\u3067\u3001\u9ad8\u5ea6\u306a\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u80fd\u529b\u3068\u5206\u6790\u7684\u601d\u8003\u306e\u4e21\u65b9\u3092\u8981\u6c42\u3059\u308b\u3002\u3053\u308c\u3089\u306e\u4f5c\u696d\u306f\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u5f97\u610f\u3068\u3059\u308b\u9818\u57df\u3068\u5927\u304d\u304f\u91cd\u306a\u3063\u3066\u304a\u308a\u3001\u81ea\u52d5\u5316\u306b\u3088\u308b\u52b9\u7387\u5316\u306e\u4f59\u5730\u304c\u5927\u304d\u3044\u3068\u8003\u3048\u3089\u308c\u308b\u3002<\/p>\n\n\n\n

\u672c\u30d6\u30ed\u30b0\u3067\u306f\u3001\u6a5f\u68b0\u5b66\u7fd2\u30b3\u30f3\u30da\u30c6\u30a3\u30b7\u30e7\u30f3\u306e\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u3067\u3042\u308bKaggle\u306e\u5165\u9580\u7684\u8ab2\u984c\u300cTitanic:Machine Learning from Disaster\u300d\u3092\u5bfe\u8c61\u306b\u3001Claude Code\u3092\u7528\u3044\u305f\u81ea\u52d5\u6539\u5584\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u3092\u69cb\u7bc9\u3057\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u53cd\u5fa9\u7684\u306a\u8a66\u884c\u3092\u901a\u3058\u3066\u3069\u306e\u7a0b\u5ea6\u306e\u4e88\u6e2c\u7cbe\u5ea6\u3092\u9054\u6210\u3067\u304d\u308b\u304b\u3092\u691c\u8a3c\u3059\u308b\u3002<\/p>\n\n\n\n

\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u69cb\u7bc9\u306e\u30a2\u30fc\u30ad\u30c6\u30af\u30c1\u30e3<\/h2>\n\n\n\n

\u5168\u4f53\u6982\u8981<\/h4>\n\n\n\n

\u672c\u30b7\u30b9\u30c6\u30e0\u306f\u3001Python\u88fd\u306e\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u30b9\u30af\u30ea\u30d7\u30c8\uff08PipelineAgent.py<\/a>\uff09\u304c4\u3064\u306eClaudeCode\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3092\u9806\u6b21\u547c\u3073\u51fa\u3059\u69cb\u6210\u3092\u3068\u308b\u3002\u5404\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306fClaude Code CLI\u3092\u975e\u5bfe\u8a71\u30e2\u30fc\u30c9\uff08claude -p<\/code>\uff09\u3067\u30b5\u30d6\u30d7\u30ed\u30bb\u30b9\u3068\u3057\u3066\u5b9f\u884c\u3057\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u9593\u306e\u60c5\u5831\u5171\u6709\u306f\u30d5\u30a1\u30a4\u30eb\u30b7\u30b9\u30c6\u30e0\u4e0a\u306eMarkdown\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u3092\u4ecb\u3057\u3066\u884c\u3046\u3002<\/p>\n\n\n\n

\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u306f1\u56de\u306e\u5b9f\u9a13\u3054\u3068\u306b\u72ec\u7acb\u3057\u305f\u30c7\u30a3\u30ec\u30af\u30c8\u30ea\uff08experiment_n<\/code>\uff09\u3092\u4f5c\u6210\u3057\u3001\u8907\u6570\u56de\u306e\u53cd\u5fa9\u5b9f\u884c\u306b\u3088\u3063\u3066\u30b9\u30b3\u30a2\u306e\u81ea\u52d5\u6539\u5584\u3092\u76ee\u6307\u3059\u8a2d\u8a08\u3068\u306a\u3063\u3066\u3044\u308b\u3002<\/p>\n\n\n\n

\u51e6\u7406\u30d5\u30ed\u30fc<\/h4>\n\n\n\n
    \n
  1. ClaudeCode1<\/strong>\uff1a\u904e\u53bb\u306e\u5b9f\u9a13\u7d50\u679c\uff08result.md<\/a>\uff09\u3068\u30b3\u30f3\u30da\u4ed5\u69d8\uff08reference.md<\/a>\uff09\u3092\u53c2\u7167\u3057\u3001EDA\u8a08\u753b\uff08plan_EDA.md<\/a>\uff09\u3068\u5206\u6790\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3059\u308b<\/li>\n\n\n\n
  2. \u30ed\u30fc\u30ab\u30eb\u5b9f\u884c<\/strong>\uff1aCC1\u304c\u751f\u6210\u3057\u305fEDA\u30b3\u30fc\u30c9\u3092\u30ed\u30fc\u30ab\u30eb\u74b0\u5883\u3067\u5b9f\u884c\u3057\u3001\u51fa\u529b\u7d50\u679c\u3092\u53d6\u5f97\u3059\u308b<\/li>\n\n\n\n
  3. ClaudeCode2<\/strong>\uff1aEDA\u8a08\u753b\u3068\u5b9f\u884c\u7d50\u679c\u3092\u53d7\u3051\u53d6\u308a\u3001\u5206\u6790\u5185\u5bb9\u3092EDA.md<\/a>\u3068\u3057\u3066\u307e\u3068\u3081\u308b<\/li>\n\n\n\n
  4. ClaudeCode3<\/strong>\uff1a\u30b3\u30f3\u30da\u4ed5\u69d8\u3068EDA\u7d50\u679c\u3092\u53c2\u7167\u3057\u3001\u30e2\u30c7\u30eb\u69cb\u7bc9\u8a08\u753b\uff08plan_model.md<\/a>\uff09\u3068\u5b66\u7fd2\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3059\u308b<\/li>\n\n\n\n
  5. \u30ed\u30fc\u30ab\u30eb\u5b9f\u884c<\/strong>\uff1aCC3\u304c\u751f\u6210\u3057\u305f\u30e2\u30c7\u30eb\u30b3\u30fc\u30c9\u3092\u30ed\u30fc\u30ab\u30eb\u3067\u5b9f\u884c\u3057\u3001CV\u30b9\u30b3\u30a2\u3068\u63d0\u51fa\u30d5\u30a1\u30a4\u30eb\u3092\u751f\u6210\u3059\u308b<\/li>\n\n\n\n
  6. ClaudeCode4<\/strong>\uff1a\u30e2\u30c7\u30eb\u8a08\u753b\u3068\u5b9f\u884c\u7d50\u679c\u3092\u53d7\u3051\u53d6\u308a\u3001\u7cbe\u5ea6\u7d50\u679c\u3068\u8003\u5bdf\u3092result.md<\/a>\u3068\u3057\u3066\u307e\u3068\u3081\u308b<\/li>\n<\/ol>\n\n\n\n

    \u3053\u306e\u4e00\u9023\u306e\u30d5\u30ed\u30fc\u3092\u4efb\u610f\u306e\u56de\u6570\u7e70\u308a\u8fd4\u3059\u3053\u3068\u3067\u3001CC1\u304c\u904e\u53bb\u306e\u7d50\u679c\u3092\u53c2\u7167\u3057\u306a\u304c\u3089\u5206\u6790\u65b9\u91dd\u3092\u6539\u5584\u3057\u3001\u30b9\u30b3\u30a2\u306e\u5411\u4e0a\u3092\u56f3\u308b\u3002<\/p>\n\n\n\n

    \u30c7\u30a3\u30ec\u30af\u30c8\u30ea\u69cb\u6210<\/h2>\n\n\n\n

    \u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u306f competition<\/code>\uff08\u30b3\u30f3\u30da\u95a2\u9023\u30d5\u30a1\u30a4\u30eb\uff09\u3068 improvement<\/code>\uff08\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u30fb\u5b9f\u9a13\u7d50\u679c\uff09\u306e2\u3064\u306e\u30c7\u30a3\u30ec\u30af\u30c8\u30ea\u306b\u5206\u304b\u308c\u308b\u3002<\/p>\n\n\n\n

    Auto_Improvement\/\n\u251c\u2500\u2500 competition\/\n\u2502   \u251c\u2500\u2500 data\/\n\u2502   \u2502   \u251c\u2500\u2500 train.csv        # \u5b66\u7fd2\u30c7\u30fc\u30bf\n\u2502   \u2502   \u2514\u2500\u2500 test.csv         # \u63a8\u8ad6\u30c7\u30fc\u30bf\n\u2502   \u2514\u2500\u2500 reference.md         # \u30b3\u30f3\u30da\u4ed5\u69d8\u30fb\u8a55\u4fa1\u65b9\u6cd5\u306e\u5b9a\u7fa9\n\u2502\n\u2514\u2500\u2500 improvement\/\n    \u251c\u2500\u2500 PipelineAgent.py     # \u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u5b9f\u884c\u30b9\u30af\u30ea\u30d7\u30c8\n    \u2514\u2500\u2500 agent_results\/\n        \u2514\u2500\u2500 experiment_n\/    # \u5b9f\u9a13\u3054\u3068\u306b\u81ea\u52d5\u63a1\u756a\n            \u251c\u2500\u2500 middle_data\/\n            \u2502   \u2514\u2500\u2500 middle_data.csv   # EDA\u3067\u751f\u6210\u3057\u305f\u4e2d\u9593\u30c7\u30fc\u30bf\n            \u2514\u2500\u2500 Result\/\n                \u251c\u2500\u2500 plan_EDA.md       # EDA\u8a08\u753b\uff08CC1\u304c\u751f\u6210\uff09\n                \u251c\u2500\u2500 eda_analysis.py   # EDA\u5b9f\u884c\u30b3\u30fc\u30c9\uff08CC1\u304c\u751f\u6210\uff09\n                \u251c\u2500\u2500 EDA.md            # EDA\u7d50\u679c\u307e\u3068\u3081\uff08CC2\u304c\u751f\u6210\uff09\n                \u251c\u2500\u2500 plan_model.md     # \u30e2\u30c7\u30eb\u8a08\u753b\uff08CC3\u304c\u751f\u6210\uff09\n                \u251c\u2500\u2500 model_train.py    # \u30e2\u30c7\u30eb\u5b66\u7fd2\u30b3\u30fc\u30c9\uff08CC3\u304c\u751f\u6210\uff09\n                \u251c\u2500\u2500 result.md         # \u5b9f\u9a13\u7d50\u679c\u307e\u3068\u3081\uff08CC4\u304c\u751f\u6210\uff09\n                \u2514\u2500\u2500 submission.csv    # Kaggle\u63d0\u51fa\u30d5\u30a1\u30a4\u30eb\n<\/code><\/pre>\n\n\n\n
    reference.md<\/code> \u306b\u306f\u30b3\u30f3\u30da\u6982\u8981\u30fb\u30c7\u30fc\u30bf\u306e\u30ab\u30e9\u30e0\u5b9a\u7fa9\u30fb\u8a55\u4fa1\u6307\u6a19\uff08Accuracy\uff09\u30fb\u5b66\u7fd2\u6642\u306e\u8a55\u4fa1\u65b9\u6cd5\uff08Stratified 5-Fold CV\uff09\u3092\u8a18\u8f09\u3057\u3066\u304a\u308a\u3001\u5404\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u30a8\u30e9\u30fc\u306e\u306a\u3044\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3059\u308b\u305f\u3081\u306e\u4ed5\u69d8\u66f8\u3068\u3057\u3066\u6a5f\u80fd\u3059\u308b\u3002\u5b9f\u9a13\u306f experiment_1<\/code>, experiment_2<\/code> \u2026 \u3068\u81ea\u52d5\u63a1\u756a\u3055\u308c\u3001\u904e\u53bb\u306e\u5168 result.md<\/code> \u304c\u6b21\u306e\u5b9f\u9a13\u306eCC1\u3078\u306e\u5165\u529b\u3068\u306a\u308b\u3002<\/p>\n\n\n\n
    \u8a73\u7d30\u30b3\u30fc\u30c9<\/h4>\n\n\n\n
    \u8a73\u7d30\u306a\u30b3\u30fc\u30c9\u306f\u4ee5\u4e0b\u306eGithub\u306b\u8f09\u305b\u308b\u3053\u3068\u3068\u3059\u308b\u3002\u306a\u304a\u3001\u4e2d\u9593\u30c7\u30fc\u30bf\u3092\u542b\u3081\u30c7\u30fc\u30bf\u306fGithub\u306b\u8f09\u305b\u3066\u3044\u306a\u3044\u3002<\/p>\n\n\n\n
    https:\/\/github.com\/keisuke-fbr\/auto-improvement<\/a><\/p>\n\n\n\n
    \u5b9f\u9a13\u5185\u5bb9<\/h2>\n\n\n\n
    \u5b9f\u9a13\u8a2d\u5b9a<\/h4>\n\n\n\n
    \u672c\u5b9f\u9a13\u3067\u306f\u3001Kaggle\u300cTitanic: Machine Learning from Disaster\u300d\u3092\u5bfe\u8c61\u306b\u3001\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u306e\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u56de\u6570\u304c\u6700\u7d42\u7684\u306a\u4e88\u6e2c\u7cbe\u5ea6\u306b\u4e0e\u3048\u308b\u5f71\u97ff\u3092\u691c\u8a3c\u3059\u308b\u3002\u5177\u4f53\u7684\u306b\u306f\u4ee5\u4e0b\u306e2\u6761\u4ef6\u3067\u5b9f\u9a13\u3092\u5b9f\u65bd\u3057\u3001CV\u30b9\u30b3\u30a2\uff08Stratified 5-Fold Accuracy\uff09\u3092\u6bd4\u8f03\u3059\u308b\u3002<\/p>\n\n\n\n
    \u6761\u4ef6<\/th>\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u6570<\/th>\u6982\u8981<\/th><\/tr><\/thead>
    \u6761\u4ef6A<\/td>1\u56de<\/td>\u521d\u56de\u306e\u5206\u6790\u30fb\u30e2\u30c7\u30eb\u69cb\u7bc9\u7d50\u679c\u3092\u305d\u306e\u307e\u307e\u8a55\u4fa1<\/td><\/tr>
    \u6761\u4ef6B<\/td>3\u56de<\/td>\u904e\u53bb\u306e\u7d50\u679c\u3092\u53c2\u7167\u3057\u306a\u304c\u3089\u5206\u6790\u3068\u30e2\u30c7\u30eb\u3092\u53cd\u5fa9\u6539\u5584<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
    \u8a55\u4fa1\u6307\u6a19<\/h4>\n\n\n\n
      \n
    • \u305d\u308c\u305e\u308c\u306e\u5b9f\u884c\u5f8c\u3001\u751f\u6210\u3055\u308c\u308bsubmission.csv\u306ePublib\u30b9\u30b3\u30a2\u3067\u8a55\u4fa1\u3092\u884c\u3046\u3002<\/li>\n<\/ul>\n\n\n\n
      \u5b9f\u9a13\u306e\u7740\u773c\u70b9<\/h4>\n\n\n\n
      \u6761\u4ef6B\u3067\u306f\u3001CC1\u304c\u904e\u53bb\u306e result.md<\/code> \u3092\u3059\u3079\u3066\u53c2\u7167\u3057\u305f\u4e0a\u3067EDA\u65b9\u91dd\u3092\u6c7a\u5b9a\u3059\u308b\u305f\u3081\u3001\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u3092\u91cd\u306d\u308b\u3054\u3068\u306b\u7279\u5fb4\u91cf\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\u3084\u30e2\u30c7\u30eb\u9078\u629e\u304c\u6539\u5584\u3055\u308c\u308b\u3053\u3068\u304c\u671f\u5f85\u3055\u308c\u308b\u3002\u672c\u5b9f\u9a13\u3067\u306f\u305d\u306e\u52b9\u679c\u3092\u5b9a\u91cf\u7684\u306b\u78ba\u8a8d\u3059\u308b\u3002<\/p>\n\n\n\n
      \n\n\n\n
      \u5b9f\u9a13\u7d50\u679c<\/h2>\n\n\n\n
      \u6761\u4ef6A\uff081\u56de\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\uff09\u306e\u7d50\u679c<\/h3>\n\n\n\n
      \u6700\u7d42\u8a55\u4fa1\u7d50\u679c<\/h4>\n\n\n\n
      \u6307\u6a19<\/th>\u30b9\u30b3\u30a2<\/th><\/tr><\/thead>
      Kaggle Public Score (Accuracy)<\/td>0.76794<\/strong><\/td><\/tr>
      CV Accuracy\uff08Stratified 5-Fold \u5e73\u5747\uff09<\/td>0.8350 \u00b1 0.0171<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u524d\u51e6\u7406\u5185\u5bb9<\/h4>\n\n\n\n
      \u6b20\u640d\u5024\u88dc\u5b8c<\/strong><\/p>\n\n\n\n
      \u30ab\u30e9\u30e0<\/th>\u88dc\u5b8c\u65b9\u6cd5<\/th><\/tr><\/thead>
      Age<\/td>Title\u30b0\u30eb\u30fc\u30d7\u5225\u4e2d\u592e\u5024\uff08\u30d5\u30a9\u30fc\u30eb\u30d0\u30c3\u30af\uff1a\u5168\u4f53\u4e2d\u592e\u5024\uff09<\/td><\/tr>
      Fare<\/td>train\u306e\u4e2d\u592e\u5024\uff08test 1\u4ef6\u306e\u307f\u6b20\u640d\uff09<\/td><\/tr>
      Embarked<\/td>\u6700\u983b\u5024\uff08”S”\uff09<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u7279\u5fb4\u91cf\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\uff08\u8a0814\u6b21\u5143\uff09<\/strong><\/p>\n\n\n\n
      \u7279\u5fb4\u91cf<\/th>\u5185\u5bb9<\/th><\/tr><\/thead>
      Title<\/td>Name\u304b\u3089\u656c\u79f0\u62bd\u51fa\uff08Mr \/ Mrs \/ Miss \/ Master \/ Rare\uff09<\/td><\/tr>
      FamilySize<\/td>SibSp + Parch + 1<\/td><\/tr>
      IsAlone<\/td>FamilySize == 1 \u306a\u3089 1\u3001\u305d\u308c\u4ee5\u5916\u306f 0<\/td><\/tr>
      HasCabin<\/td>Cabin\u975e\u6b20\u640d\u306a\u3089 1\u3001\u6b20\u640d\u306a\u3089 0<\/td><\/tr>
      LogFare<\/td>log1p(Fare)<\/td><\/tr>
      AgeBin<\/td>\u5e74\u9f62\u30925\u533a\u5206\u306b\u30d3\u30cb\u30f3\u30b0<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u30ab\u30c6\u30b4\u30ea\u5909\u6570\uff08Sex\u30fbTitle\u30fbAgeBin\uff09\u306fLabel Encoding\u3001Embark\u306fOne-Hot Encoding\u3092\u9069\u7528\u3002<\/p>\n\n\n\n
      \u4f7f\u7528\u30e2\u30c7\u30eb<\/h4>\n\n\n\n
      RandomForestClassifier \u3068 GradientBoostingClassifier \u306e\u30a2\u30f3\u30b5\u30f3\u30d6\u30eb\uff08VotingClassifier\u30fbsoft voting\uff09\u3002<\/p>\n\n\n\n
      \u30e2\u30c7\u30eb<\/th>\u4e3b\u8981\u30d1\u30e9\u30e1\u30fc\u30bf<\/th><\/tr><\/thead>
      RandomForestClassifier<\/td>n_estimators=500, max_depth=8<\/td><\/tr>
      GradientBoostingClassifier<\/td>n_estimators=200, learning_rate=0.05, max_depth=4<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u6761\u4ef6B\uff083\u56de\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\uff09\u306e\u7d50\u679c<\/h3>\n\n\n\n
      \n
      \u4ee5\u4e0b\u306e\u524d\u51e6\u7406\u30fb\u30e2\u30c7\u30eb\u60c5\u5831\u306f\u30013\u56de\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u306e\u5185\u6700\u7d42\u5b9f\u9a13\u3067\u3042\u308b experiment_3<\/strong> \u306e\u7d50\u679c\u306b\u57fa\u3065\u3044\u3066\u3044\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n
      \u6700\u7d42\u8a55\u4fa1\u7d50\u679c<\/h4>\n\n\n\n
      \u6307\u6a19<\/th>\u30b9\u30b3\u30a2<\/th><\/tr><\/thead>
      Kaggle Public Score (Accuracy)<\/td>0.77990<\/strong><\/td><\/tr>
      CV Accuracy\uff08Stratified 5-Fold \u5e73\u5747\uff09<\/td>0.8283 \u00b1 0.0151<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u524d\u51e6\u7406\u5185\u5bb9<\/h4>\n\n\n\n
      \u6b20\u640d\u5024\u88dc\u5b8c<\/strong><\/p>\n\n\n\n
      \u30ab\u30e9\u30e0<\/th>\u88dc\u5b8c\u65b9\u6cd5<\/th><\/tr><\/thead>
      Age<\/td>Title \u00d7 Pclass\u30b0\u30eb\u30fc\u30d7\u5225\u4e2d\u592e\u5024\uff08train\u3067\u8a08\u7b97\u3001test\u306b\u9069\u7528\uff09<\/td><\/tr>
      Fare<\/td>train\u306e\u4e2d\u592e\u5024<\/td><\/tr>
      Embarked<\/td>\u6700\u983b\u5024\uff08”S”\uff09<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u7279\u5fb4\u91cf\u30a8\u30f3\u30b8\u30cb\u30a2\u30ea\u30f3\u30b0\uff08\u8a0817\u6b21\u5143\uff09<\/strong><\/p>\n\n\n\n
      \u6761\u4ef6A\u306e14\u6b21\u5143\u304b\u3089\u3055\u3089\u306b\u4ee5\u4e0b3\u7279\u5fb4\u91cf\u3092\u8ffd\u52a0\u3057\u305f\u3002<\/p>\n\n\n\n
      \u7279\u5fb4\u91cf<\/th>\u5185\u5bb9<\/th><\/tr><\/thead>
      FarePerPerson<\/td>Fare \/ \u540c\u4e00\u30c1\u30b1\u30c3\u30c8\u306e\u4eba\u6570<\/td><\/tr>
      NameLength<\/td>Name\u306e\u6587\u5b57\u6570<\/td><\/tr>
      AgeClass<\/td>Age \u00d7 Pclass<\/td><\/tr>
      TicketPrefix<\/td>\u30c1\u30b1\u30c3\u30c8\u756a\u53f7\u306e\u82f1\u5b57\u30d7\u30ec\u30d5\u30a3\u30c3\u30af\u30b9<\/td><\/tr>
      TicketGroupSize<\/td>\u540c\u4e00\u30c1\u30b1\u30c3\u30c8\u756a\u53f7\u306e\u4eba\u6570<\/td><\/tr>
      IsChild<\/td>Age < 12 \u306e\u30d5\u30e9\u30b0<\/td><\/tr>
      TicketGroupSurvRate<\/td>\u30c1\u30b1\u30c3\u30c8\u30b0\u30eb\u30fc\u30d7\u306e\u751f\u5b58\u7387\uff08middle_data\u304b\u3089\u53d6\u5f97\uff09<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u4f7f\u7528\u30e2\u30c7\u30eb<\/h4>\n\n\n\n
      RandomForest\u5358\u4f53\u3067\u306e\u63a8\u8ad6\u3068\u306a\u3063\u305f\u3002<\/p>\n\n\n\n
      \u30e2\u30c7\u30eb<\/th>\u4e3b\u8981\u30d1\u30e9\u30e1\u30fc\u30bf<\/th><\/tr><\/thead>
      RandomForestClassifier<\/td>n_estimators=300, max_depth=8, min_samples_leaf=5<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \n\n\n\n
      \u307e\u3068\u3081<\/h2>\n\n\n\n
      \u672c\u5b9f\u9a13\u3067\u306f\u3001Claude Code\u3092\u7528\u3044\u305f\u81ea\u52d5\u6539\u5584\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u306b\u3088\u3063\u3066\u3001Kaggle Titanic\u30b3\u30f3\u30da\u306b\u304a\u3051\u308b\u4e88\u6e2c\u7cbe\u5ea6\u306e\u5909\u5316\u3092\u691c\u8a3c\u3057\u305f\u3002<\/p>\n\n\n\n
      \u6761\u4ef6<\/th>\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u6570<\/th>Kaggle Score<\/th><\/tr><\/thead>
      \u6761\u4ef6A<\/td>1\u56de<\/td>0.76794<\/td><\/tr>
      \u6761\u4ef6B<\/td>3\u56de<\/td>0.77990<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
      \u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u3092\u53cd\u5fa9\u3059\u308b\u3053\u3068\u3067\u30b9\u30b3\u30a2\u306e\u5411\u4e0a\u306f\u78ba\u8a8d\u3067\u304d\u305f\u3082\u306e\u306e\u3001\u4e0a\u3052\u5e45\u306f0.012\u7a0b\u5ea6\u306b\u3068\u3069\u307e\u308a\u3001\u5927\u304d\u306a\u6539\u5584\u306f\u5f97\u3089\u308c\u306a\u304b\u3063\u305f\u3002\u3053\u306e\u8981\u56e0\u3068\u3057\u3066\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u6301\u3064\u60c5\u5831\u304c\u30ed\u30fc\u30ab\u30eb\u306e\u30c7\u30fc\u30bf\u3068\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u306e\u307f\u306b\u9650\u3089\u308c\u3066\u304a\u308a\u3001Web\u4e0a\u306e\u4e0a\u4f4d\u89e3\u6cd5\u3084\u6700\u65b0\u7814\u7a76\u306a\u3069\u306e\u5916\u90e8\u77e5\u8b58\u3092\u6d3b\u7528\u3067\u304d\u3066\u3044\u306a\u3044\u70b9\u304c\u8003\u3048\u3089\u308c\u308b\u3002\u4eca\u5f8c\u306f\u3001Web\u691c\u7d22\u6a5f\u80fd\u3092\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306b\u6301\u305f\u305b\u308b\u3053\u3068\u3084\u3001\u3055\u3089\u306a\u308b\u5b9f\u884c\u56de\u6570\u306e\u5897\u52a0\u306b\u3088\u3063\u3066\u3001\u30b9\u30b3\u30a2\u306e\u5927\u5e45\u306a\u6539\u5584\u304c\u898b\u8fbc\u3081\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u3002<\/p>\n\n\n\n\n","protected":false},"excerpt":{"rendered":"
      \u306f\u3058\u3081\u306b \u8fd1\u5e74\u3001\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff08LLM\uff09\u3092\u6d3b\u7528\u3057\u305f\u81ea\u5f8b\u7684\u306a\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u6025\u901f\u306b\u666e\u53ca\u3057\u3066\u304a\u308a\u3001\u7279\u306b\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u958b\u767a\u306e\u9818\u57df\u306b\u304a\u3044\u3066\u30b3\u30fc\u30c9\u751f\u6210\u30fb\u4fee\u6b63\u30fb\u5b9f\u884c\u3092\u81ea\u52d5\u5316\u3059\u308b\u30c4\u30fc\u30eb\u3068\u3057\u3066\u5e83\u304f\u6d3b\u7528\u3055\u308c\u3066\u3044\u308b\u3002\u305d\u306e\u4ee3\u8868\u4f8b\u3068\u3057\u3066\u3001Anth […]<\/p>\n","protected":false},"author":44,"featured_media":8467,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","swell_btn_cv_data":"","footnotes":"","_wp_rev_ctl_limit":""},"categories":[1],"tags":[96,984,1265,39,419,651],"class_list":["post-8466","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized","tag-ai","tag-claude","tag-claude-code","tag-39","tag-419"],"_links":{"self":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts\/8466","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/users\/44"}],"replies":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/comments?post=8466"}],"version-history":[{"count":1,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts\/8466\/revisions"}],"predecessor-version":[{"id":8468,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/posts\/8466\/revisions\/8468"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/media\/8467"}],"wp:attachment":[{"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/media?parent=8466"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/categories?post=8466"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/since2020.jp\/media\/wp-json\/wp\/v2\/tags?post=8466"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}