{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np\nimport pandas as pd\n\nfrom lightgbm import LGBMClassifier\nfrom sklearn.ensemble import ExtraTreesClassifier\nfrom sklearn import model_selection\nfrom sklearn.metrics import roc_auc_score","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"data = pd.concat([\n       pd.read_csv(\"../input/pageviews/pageviews.csv\", parse_dates=[\"FEC_EVENT\"]),\n       pd.read_csv(\"../input/pageviews_complemento/pageviews_complemento.csv\", parse_dates=[\"FEC_EVENT\"])\n])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test = []\nfor c in data.drop([\"USER_ID\", \"FEC_EVENT\"], axis=1).columns:\n    print(\"haciendo\", c)\n    temp = pd.crosstab(data.USER_ID, data[c])\n    temp.columns = [c + \"_\" + str(v) for v in temp.columns]\n    X_test.append(temp.apply(lambda x: x / x.sum(), axis=1))\nX_test = pd.concat(X_test, axis=1)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"data = data[data.FEC_EVENT.dt.month < 10]\nX_train = []\nfor c in data.drop([\"USER_ID\", \"FEC_EVENT\"], axis=1).columns:\n    print(\"haciendo\", c)\n    temp = pd.crosstab(data.USER_ID, data[c])\n    temp.columns = [c + \"_\" + str(v) for v in temp.columns]\n    X_train.append(temp.apply(lambda x: x / x.sum(), axis=1))\nX_train = pd.concat(X_train, axis=1)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"features = list(set(X_train.columns).intersection(set(X_test.columns)))\nX_train = X_train[features]\nX_test = X_test[features]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"y_prev = pd.read_csv(\"../input/conversiones/conversiones.csv\")\ny_train = pd.Series(0, index=X_train.index)\nidx = set(y_prev[y_prev.mes >= 10].USER_ID.unique()).intersection(\n        set(X_train.index))\ny_train.loc[list(idx)] = 1","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"ExtraTreesClassifier?","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"candidatos = [{\n    \"learner\": LGBMClassifier,\n    \"param_grid\": model_selection.ParameterGrid({\n           \"n_estimators\": [10000],\n           \"num_leaves\": [200, 30, 40],\n           \"max_depth\": [-1, 3, 5, 10],\n           \"min_child_samples\": [100, 20],\n    }),\n    \"train_params\": {\n        \"early_stopping_rounds\": 10,\n        \"eval_metric\": \"auc\", \n        \"eval_set\": True,\n        \"verbose\": 25\n    } \n}, {\n    \"learner\": ExtraTreesClassifier,\n    \"param_grid\": model_selection.ParameterGrid({\n           \"n_estimators\": [1000],\n           \"min_samples_leaf\": [10, 20, 40, 100],\n    })\n}]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"folds = list(model_selection.KFold(n_splits=3, shuffle=True).split(X_train))\nres = []\nbestRes = 0\nbestProbs = y_train\nfor candidate in candidatos:\n    for params in candidate[\"param_grid\"]:\n        trainParams = candidate.get(\"train_params\", {})\n        valid_probs = []\n        test_probs = []\n        for i, (train_idx, valid_idx) in enumerate(folds):\n            Xt = X_train.iloc[train_idx]\n            yt = y_train.loc[X_train.index].iloc[train_idx]\n\n            Xv = X_train.iloc[valid_idx]\n            yv = y_train.loc[X_train.index].iloc[valid_idx]\n            if \"eval_set\" in trainParams:\n                trainParams[\"eval_set\"] = [(Xt, yt), (Xv, yv)]\n            learner = candidate[\"learner\"](**params)\n            learner.fit(Xt, yt, **trainParams)\n\n            valid_probs.append(pd.Series(learner.predict_proba(Xv)[:, -1], index=Xv.index, name=\"SCORE\"))\n            test_probs.append(pd.Series(learner.predict_proba(X_test)[:, -1],\n                                        index=X_test.index, name=\"fold_\" + str(i)))\n\n        valid_probs = pd.concat(valid_probs)\n        cres = roc_auc_score(y_train, valid_probs.loc[y_train.index])\n        cols = [\"learner\"]\n        vals = [candidate[\"learner\"].__name__]\n        for p in params:\n            cols.append(\"param_\" + p)\n            vals.append(params[p])\n        cols.append(\"score\")\n        vals.append(cres)\n        res.append(pd.DataFrame([vals], columns=cols))\n        print(\"*\"*10)\n        print(pd.concat(res))\n        print(\"*\"*10)\n        if cres > bestRes:\n            test_probs = pd.concat(test_probs, axis=1).mean(axis=1)\n            test_probs.index.name=\"USER_ID\"\n            test_probs.name=\"SCORE\"\n            bestProbs = test_probs\n            bestRes = cres\n\nbestProbs.to_csv(\"benchmark.csv\", header=True)\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"pd.concat(res).sort_values(\"score\")","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}