\nUnfortunately, I did not select the high scoring code results for the B-list。
\n
![\"another](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20613639%2Fa055cda6b1b9234e5826d5e153decc5f%2F.png?generation=1716858659428039&alt=media\")
\n
![\"another](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20613639%2F0aaf100debd7b16bde01108316fae76b%2F2.png?generation=1716858835318399&alt=media\")
Thank you to the organizers and open-source contributors, you have brought me a lot of benefits。
\nUnfortunately, I did not select the high scoring code results for the B-list。
\n
\n
Here, I would like to publicly disclose this method, which is probably the one chosen by the front row to maintain a stable ranking。Weeknum can be sorted through the dpdmaxdateyear_596T、dpdmaxdatemonth_89T、overdueamountmaxdateyear_2T、overdueamountmaxdatemonth_365T；
\nAdd the following post-processing on the basis of open-source 0.584 or 0.585 code，You can get a B-list score similar to the gold medal zone in the picture above；
def (regex_path, depth=None):\n chunks = []\n for path in ((regex_path)):\n df = pl.(path)\n df = df.(Pipeline.set_table_dtypes)\n df = df.([, , ,\n ,,])\n df = df.(df[] == )\n df = df.([])\n chunks.(df)\n df = pl.(chunks, how=)\n df = df.(subset=[])\n return df\n\ndf_base = (TEST_DIR / )\ntmp = (TEST_DIR / )\ntmp = df_base.(tmp, how=, on=)\ntmp = tmp.([, , ,\n ,]).()\n\ndf_test = df_test.(columns=[])\ndf_test = df_test.()\n\ny_pred = pd.(model.(df_test)[:, ], index=df_test.index)\n\ndf_subm = pd.(ROOT / )\ndf_subm = df_subm.()\ndf_subm[] = y_pred\n\ndf_subm = df_subm.(tmp,how=,on=[])\n\nyear_596T_ratio = df_subm[].().()/ (df_subm)\nyear_2T_ratio = df_subm[].().()/ (df_subm)\n\nif year_596T_ratio <= year_2T_ratio:\n df_subm = df_subm.(by=[, ])\nelse:\n df_subm = df_subm.(by=[, ])\n\ndf_subm = df_subm.(drop=True)\nscore_col_index = df_subm.columns.()\ncut_off_index = ((df_subm) * )\ndf_subm.iloc[:cut_off_index, score_col_index] = (df_subm.iloc[:cut_off_index, score_col_index] - ).()\ndf_subm= df_subm.([,,\n ,],axis=)\ndf_subm = df_subm.()\ndf_subm.()\nAlthough I didn't have a stable CV due to the influence of metrics, participating in this competition was more of a gain for me;
\nI will tidy up my mood and embark on the next kaggle journey. Thank you all, maybe we can meet again on Automated Essay Scoring 2.0.
Thank you to the organizers and open-source contributors, you have brought me a lot of benefits。
\nUnfortunately, I did not select the high scoring code results for the B-list。
\n
\n
Here, I would like to publicly disclose this method, which is probably the one chosen by the front row to maintain a stable ranking。Weeknum can be sorted through the dpdmaxdateyear_596T、dpdmaxdatemonth_89T、overdueamountmaxdateyear_2T、overdueamountmaxdatemonth_365T；
\nAdd the following post-processing on the basis of open-source 0.584 or 0.585 code，You can get a B-list score similar to the gold medal zone in the picture above；
def (regex_path, depth=None):\n chunks = []\n for path in ((regex_path)):\n df = pl.(path)\n df = df.(Pipeline.set_table_dtypes)\n df = df.([, , ,\n ,,])\n df = df.(df[] == )\n df = df.([])\n chunks.(df)\n df = pl.(chunks, how=)\n df = df.(subset=[])\n return df\n\ndf_base = (TEST_DIR / )\ntmp = (TEST_DIR / )\ntmp = df_base.(tmp, how=, on=)\ntmp = tmp.([, , ,\n ,]).()\n\ndf_test = df_test.(columns=[])\ndf_test = df_test.()\n\ny_pred = pd.(model.(df_test)[:, ], index=df_test.index)\n\ndf_subm = pd.(ROOT / )\ndf_subm = df_subm.()\ndf_subm[] = y_pred\n\ndf_subm = df_subm.(tmp,how=,on=[])\n\nyear_596T_ratio = df_subm[].().()/ (df_subm)\nyear_2T_ratio = df_subm[].().()/ (df_subm)\n\nif year_596T_ratio <= year_2T_ratio:\n df_subm = df_subm.(by=[, ])\nelse:\n df_subm = df_subm.(by=[, ])\n\ndf_subm = df_subm.(drop=True)\nscore_col_index = df_subm.columns.()\ncut_off_index = ((df_subm) * )\ndf_subm.iloc[:cut_off_index, score_col_index] = (df_subm.iloc[:cut_off_index, score_col_index] - ).()\ndf_subm= df_subm.([,,\n ,],axis=)\ndf_subm = df_subm.()\ndf_subm.()\nAlthough I didn't have a stable CV due to the influence of metrics, participating in this competition was more of a gain for me;
\nI will tidy up my mood and embark on the next kaggle journey. Thank you all, maybe we can meet again on Automated Essay Scoring 2.0.
Wonderful!
", "rawMarkdown": "Wonderful!", "votes": null }, { "id": "2840179", "postDate": "05/28/2024 02:22:52", "content": "太可惜了，大佬！What a pity
", "rawMarkdown": "太可惜了，大佬！What a pity", "votes": null }, { "id": "2840194", "postDate": "05/28/2024 02:38:22", "content": "Awesome !!!
", "rawMarkdown": "Awesome !!!", "votes": null }, { "id": "2840646", "postDate": "05/28/2024 07:26:37", "content": "Love youuuuu!!!!!!!!!!! <3.
", "rawMarkdown": "Love youuuuu!!!!!!!!!!! <3.", "votes": null }, { "id": "2841655", "postDate": "05/28/2024 16:36:30", "content": "确实牛 实在太可惜了
", "rawMarkdown": "确实牛 实在太可惜了", "votes": null } ], "comments": [ { "id": 2840152, "author_name": "polsyuus", "author_url": "", "post_date": "05/28/2024 01:53:00", "content": "Wonderful!
", "votes": null, "replies": [] }, { "id": 2840179, "author_name": "wessonz", "author_url": "", "post_date": "05/28/2024 02:22:52", "content": "太可惜了，大佬！What a pity
", "votes": null, "replies": [] }, { "id": 2840194, "author_name": "nguyenosaurus", "author_url": "", "post_date": "05/28/2024 02:38:22", "content": "Awesome !!!
", "votes": null, "replies": [] }, { "id": 2840646, "author_name": "marcyyl", "author_url": "", "post_date": "05/28/2024 07:26:37", "content": "Love youuuuu!!!!!!!!!!! <3.
", "votes": null, "replies": [] }, { "id": 2841655, "author_name": "licongdu", "author_url": "", "post_date": "05/28/2024 16:36:30", "content": "确实牛 实在太可惜了
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2840124": "Thank you to the organizers and open-source contributors, you have brought me a lot of benefits。\nUnfortunately, I did not select the high scoring code results for the B-list。\n\n\n\nHere, I would like to publicly disclose this method, which is probably the one chosen by the front row to maintain a stable ranking。Weeknum can be sorted through the dpdmaxdateyear_596T、dpdmaxdatemonth_89T、overdueamountmaxdateyear_2T、overdueamountmaxdatemonth_365T；\nAdd the following post-processing on the basis of open-source 0.584 or 0.585 code，You can get a B-list score similar to the gold medal zone in the picture above；\n```\ndef read_files2(regex_path, depth=None):\n chunks = []\n for path in glob(str(regex_path)):\n df = pl.read_parquet(path)\n df = df.pipe(Pipeline.set_table_dtypes)\n df = df.select(['case_id', 'dpdmaxdateyear_596T', 'dpdmaxdatemonth_89T',\n 'overdueamountmaxdateyear_2T','overdueamountmaxdatemonth_365T','num_group1'])\n df = df.filter(df['num_group1'] == 0)\n df = df.drop(['num_group1'])\n chunks.append(df)\n df = pl.concat(chunks, how=\"vertical_relaxed\")\n df = df.unique(subset=[\"case_id\"])\n return df\n\ndf_base = read_file(TEST_DIR / \"test_base.parquet\")\ntmp = read_files2(TEST_DIR / \"test_credit_bureau_a_1_*.parquet\")\ntmp = df_base.join(tmp, how=\"left\", on=\"case_id\")\ntmp = tmp.select(['case_id', 'dpdmaxdateyear_596T', 'dpdmaxdatemonth_89T',\n 'overdueamountmaxdateyear_2T','overdueamountmaxdatemonth_365T']).to_pandas()\n\ndf_test = df_test.drop(columns=[\"WEEK_NUM\"])\ndf_test = df_test.set_index(\"case_id\")\n\ny_pred = pd.Series(model.predict_proba(df_test)[:, 1], index=df_test.index)\n\ndf_subm = pd.read_csv(ROOT / \"sample_submission.csv\")\ndf_subm = df_subm.set_index(\"case_id\")\ndf_subm[\"score\"] = y_pred\n\ndf_subm = df_subm.merge(tmp,how='left',on=['case_id'])\n\nyear_596T_ratio = df_subm['dpdmaxdateyear_596T'].isnull().sum()/ len(df_subm)\nyear_2T_ratio = df_subm['overdueamountmaxdateyear_2T'].isnull().sum()/ len(df_subm)\n\nif year_596T_ratio <= year_2T_ratio:\n df_subm = df_subm.sort_values(by=['dpdmaxdateyear_596T', 'dpdmaxdatemonth_89T'])\nelse:\n df_subm = df_subm.sort_values(by=['overdueamountmaxdateyear_2T', 'overdueamountmaxdatemonth_365T'])\n \ndf_subm = df_subm.reset_index(drop=True)\nscore_col_index = df_subm.columns.get_loc('score')\ncut_off_index = int(len(df_subm) * 0.35)\ndf_subm.iloc[:cut_off_index, score_col_index] = (df_subm.iloc[:cut_off_index, score_col_index] - 0.055).clip(0)\ndf_subm= df_subm.drop(['dpdmaxdateyear_596T','dpdmaxdatemonth_89T',\n 'overdueamountmaxdateyear_2T','overdueamountmaxdatemonth_365T'],axis=1)\ndf_subm = df_subm.set_index(\"case_id\")\ndf_subm.to_csv(\"submission.csv\")\n```\n\nAlthough I didn't have a stable CV due to the influence of metrics, participating in this competition was more of a gain for me;\nI will tidy up my mood and embark on the next kaggle journey. Thank you all, maybe we can meet again on Automated Essay Scoring 2.0.", "2840152": "Wonderful!", "2840179": "太可惜了，大佬！What a pity", "2840194": "Awesome !!!", "2840646": "Love youuuuu!!!!!!!!!!! <3.", "2841655": "确实牛 实在太可惜了" }, "source": "meta" }