{"cells":[{"metadata":{},"cell_type":"markdown","source":"With a very small dataset, lots and lots of different engineered features, vastly different CV and LB scores, and high inconsisency of LB scores with just slight changes in seed, this competition is overripe for a major shakupe. In this kernel we'll take a look at adveserial validation, and what it may portend about the shakeup."},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load in \n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport lightgbm as lgb\nfrom sklearn.model_selection import KFold\nfrom sklearn import model_selection, preprocessing, metrics\n# Input data files are available in the \"../input/\" directory.\n# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the input directory\nimport shap\nimport os\nprint(os.listdir(\"../input\"))\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport warnings\nwarnings.simplefilter(action='ignore', category=FutureWarning)\n\n# Any results you write to the current directory are saved as output.","execution_count":9,"outputs":[{"output_type":"stream","text":"['andrews-features-only', 'LANL-Earthquake-Prediction']\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"For the features I'll use a good set of engineered features. Feature angineering was done by Andrew, and I had just created a separate kernel where they can be looked at in their own right and saved:\n\nhttps://www.kaggle.com/tunguz/andrews-features-only"},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"train = pd.read_csv('../input/andrews-features-only/X_tr.csv')\ntest = pd.read_csv('../input/andrews-features-only/X_test.csv')","execution_count":2,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train.shape","execution_count":3,"outputs":[{"output_type":"execute_result","execution_count":3,"data":{"text/plain":"(4194, 138)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"test.shape","execution_count":4,"outputs":[{"output_type":"execute_result","execution_count":4,"data":{"text/plain":"(2624, 138)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"features = train.columns\ntrain['target'] = 0\ntest['target'] = 1","execution_count":5,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_test = pd.concat([train, test], axis =0)\n\ntarget = train_test['target'].values","execution_count":6,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"param = {'num_leaves': 50,\n         'min_data_in_leaf': 30, \n         'objective':'binary',\n         'max_depth': 5,\n         'learning_rate': 0.006,\n         \"min_child_samples\": 20,\n         \"boosting\": \"gbdt\",\n         \"feature_fraction\": 0.9,\n         \"bagging_freq\": 1,\n         \"bagging_fraction\": 0.9 ,\n         \"bagging_seed\": 27,\n         \"metric\": 'auc',\n         \"verbosity\": -1}\n\nfolds = KFold(n_splits=5, shuffle=True, random_state=15)\noof = np.zeros(len(train_test))\n\n\nfor fold_, (trn_idx, val_idx) in enumerate(folds.split(train_test.values, target)):\n    print(\"fold n°{}\".format(fold_))\n    trn_data = lgb.Dataset(train_test.iloc[trn_idx][features], label=target[trn_idx])\n    val_data = lgb.Dataset(train_test.iloc[val_idx][features], label=target[val_idx])\n\n    num_round = 30000\n    clf = lgb.train(param, trn_data, num_round, valid_sets = [trn_data, val_data], verbose_eval=1000, early_stopping_rounds = 1400)\n    oof[val_idx] = clf.predict(train_test.iloc[val_idx][features], num_iteration=clf.best_iteration)","execution_count":7,"outputs":[{"output_type":"stream","text":"fold n°0\nTraining until validation scores don't improve for 1400 rounds.\n[1000]\ttraining's auc: 0.940125\tvalid_1's auc: 0.832808\nEarly stopping, best iteration is:\n[35]\ttraining's auc: 0.881867\tvalid_1's auc: 0.83726\nfold n°1\nTraining until validation scores don't improve for 1400 rounds.\n[1000]\ttraining's auc: 0.938003\tvalid_1's auc: 0.846147\n[2000]\ttraining's auc: 0.980072\tvalid_1's auc: 0.844377\nEarly stopping, best iteration is:\n[827]\ttraining's auc: 0.927985\tvalid_1's auc: 0.846948\nfold n°2\nTraining until validation scores don't improve for 1400 rounds.\n[1000]\ttraining's auc: 0.935839\tvalid_1's auc: 0.854196\n[2000]\ttraining's auc: 0.978081\tvalid_1's auc: 0.856872\n[3000]\ttraining's auc: 0.995682\tvalid_1's auc: 0.857299\n[4000]\ttraining's auc: 0.999605\tvalid_1's auc: 0.857354\nEarly stopping, best iteration is:\n[3375]\ttraining's auc: 0.998053\tvalid_1's auc: 0.857614\nfold n°3\nTraining until validation scores don't improve for 1400 rounds.\n[1000]\ttraining's auc: 0.936003\tvalid_1's auc: 0.856437\n[2000]\ttraining's auc: 0.976567\tvalid_1's auc: 0.854476\nEarly stopping, best iteration is:\n[709]\ttraining's auc: 0.918229\tvalid_1's auc: 0.857096\nfold n°4\nTraining until validation scores don't improve for 1400 rounds.\n[1000]\ttraining's auc: 0.936734\tvalid_1's auc: 0.8528\n[2000]\ttraining's auc: 0.978637\tvalid_1's auc: 0.851207\nEarly stopping, best iteration is:\n[1161]\ttraining's auc: 0.944924\tvalid_1's auc: 0.852859\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"Whoa, that's a pretty significant AUC! At AUC of 0.85 there is a very significant difference between the train and test sets, and a very very good chance of a major shakeup ...\n\nLet's look now at the top 20 \"adversarial\" features."},{"metadata":{"trusted":true},"cell_type":"code","source":"feature_imp = pd.DataFrame(sorted(zip(clf.feature_importance(),train.columns)), columns=['Value','Feature'])\n\nplt.figure(figsize=(20, 10))\nsns.barplot(x=\"Value\", y=\"Feature\", data=feature_imp.sort_values(by=\"Value\", ascending=False).head(20))\nplt.title('LightGBM Features')\nplt.tight_layout()\nplt.show()\nplt.savefig('lgbm_importances-01.png')","execution_count":27,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1440x720 with 1 Axes>","image/png":"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\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 0 Axes>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":15,"outputs":[{"output_type":"execute_result","execution_count":15,"data":{"text/plain":"(138,)"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":18,"outputs":[{"output_type":"execute_result","execution_count":18,"data":{"text/plain":"<zip at 0x7fb6f0bfef48>"},"metadata":{}}]},{"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}