{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":105399,"databundleVersionId":12733338,"sourceType":"competition"},{"sourceId":248282123,"sourceType":"kernelVersion"},{"sourceId":250405784,"sourceType":"kernelVersion"},{"sourceId":247101687,"sourceType":"kernelVersion"},{"sourceId":247177692,"sourceType":"kernelVersion"}],"dockerImageVersionId":31089,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import pandas as pd\n\n# 0.43916\ndf1 = pd.read_csv(\"/kaggle/input/catboost-ranker-baseline-flightrank-2025/submission.csv\")\n# 0.42163\ndf2 = pd.read_csv(\"/kaggle/input/lightgbm-ranker-ndcg-3/submission.csv\")\n# 0.47635\ndf3 = pd.read_csv(\"/kaggle/input/ensemble-with-polars/submission.csv\")\n# 0.48388\ndf4 = pd.read_csv(\"/kaggle/input/xgboost-ranker-with-polars/submission.csv\")","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2025-07-19T14:20:57.430282Z","iopub.execute_input":"2025-07-19T14:20:57.432891Z","iopub.status.idle":"2025-07-19T14:21:08.679729Z","shell.execute_reply.started":"2025-07-19T14:20:57.432827Z","shell.execute_reply":"2025-07-19T14:21:08.679019Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import pandas as pd\n\ndfs = [\n    df1,df2,df3,df4\n]\n\ndef rank2score(sr, eps=1e-6):\n    n = sr.max()\n    return 1.0 - (sr - 1) / (n + eps)\n\nscore_frames = []\nfor i, df in enumerate(dfs):\n    tmp = df[['Id', 'ranker_id', 'selected']].copy()\n    tmp['score'] = tmp.groupby('ranker_id')['selected'].transform(rank2score)\n    score_frames.append(tmp[['Id', 'ranker_id', 'score']].rename(columns={'score': f'score_{i}'}))\n\nmerged = score_frames[0]\nfor i in range(1, 4):\n    merged = merged.merge(score_frames[i], on=['Id', 'ranker_id'], how='left')\n\nweights = [0.1, 0.1, 0.1, 0.7]\nscore_cols = [f'score_{i}' for i in range(4)]\nw = pd.Series(weights, index=score_cols)\nmerged['score_mean'] = (merged[score_cols] * w).sum(axis=1) / w.sum()\n\ndef score2rank(s):\n    return s.rank(method='first', ascending=False).astype(int)\n\nmerged['selected'] = merged.groupby('ranker_id')['score_mean'].transform(score2rank)\n\nout = merged[['Id', 'ranker_id', 'selected']]\nout.to_csv(\"submission.csv\", index=False, float_format='%.0f')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-07-19T14:25:03.484017Z","iopub.execute_input":"2025-07-19T14:25:03.484354Z","iopub.status.idle":"2025-07-19T14:26:27.041016Z","shell.execute_reply.started":"2025-07-19T14:25:03.484331Z","shell.execute_reply":"2025-07-19T14:26:27.040187Z"}},"outputs":[],"execution_count":null}]}