{"cells":[{"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)\n\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\n\nimport os\nprint(os.listdir(\"../input\"))\n\n# Any results you write to the current directory are saved as output.","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"from scipy import stats\nfrom tqdm import tqdm\nfrom sklearn.preprocessing import StandardScaler\nimport lightgbm as lgb\nimport time","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"trainData = pd.read_csv('../input/train.csv', dtype={'acoustic_data': np.int16, 'time_to_failure': np.float64})","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"pd.options.display.precision = 15","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def calRCS(data):\n    N = len(data)\n    m = np.mean(data)\n    sigma = np.std(data)\n    s = np.cumsum(data-m) * 1.0 / (N * sigma)\n    R = np.max(s) - np.min(s)\n    return R","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"rows = 100_000\nsegments = int(np.floor(trainData.shape[0] / rows))\n\nX_train = pd.DataFrame(index=range(segments), dtype=np.float64,\n                       columns=['ave', 'std', 'max', 'min','q95','q99', 'q05','q01',\n                               'abs_max', 'abs_mean', 'abs_std', 'iqr', \n                                'q999','q001','ave10'])\ny_train = pd.DataFrame(index=range(segments), dtype=np.float64,\n                       columns=['time_to_failure'])\n\nfor segment in tqdm(range(segments)):\n    seg = trainData.iloc[segment*rows:segment*rows+rows]\n    x = seg['acoustic_data'].values\n    y = seg['time_to_failure'].values[-1]\n    \n    y_train.loc[segment, 'time_to_failure'] = y\n    \n    X_train.loc[segment, 'ave'] = x.mean()\n    X_train.loc[segment, 'std'] = x.std()\n    X_train.loc[segment, 'max'] = x.max()\n    X_train.loc[segment, 'min'] = x.min()\n    X_train.loc[segment, 'q95'] = np.quantile(x,0.95)\n    X_train.loc[segment, 'q99'] = np.quantile(x,0.99)\n    X_train.loc[segment, 'q05'] = np.quantile(x,0.05)\n    X_train.loc[segment, 'q01'] = np.quantile(x,0.01)\n    \n    X_train.loc[segment, 'abs_max'] = np.abs(x).max()\n    X_train.loc[segment, 'abs_mean'] = np.abs(x).mean()\n    X_train.loc[segment, 'abs_std'] = np.abs(x).std()  \n    X_train.loc[segment, 'iqr'] = np.subtract(*np.percentile(x, [75, 25]))\n    X_train.loc[segment, 'q999'] = np.quantile(x,0.999)\n    X_train.loc[segment, 'q001'] = np.quantile(x,0.001)\n    X_train.loc[segment, 'ave10'] = stats.trim_mean(x, 0.1)\n    X_train.loc[segment, 'skew'] = stats.skew(x)\n    X_train.loc[segment, 'av_change_abs'] = np.mean(np.diff(x))\n    X_train.loc[segment, 'av_change_rate'] = np.mean(np.nonzero((np.diff(x) / x[:-1]))[0])\n\n# Extra calculated time series features\n    X_train.loc[segment, 'rcs'] = calRCS(x)\n    X_train.loc[segment, 'amplitude'] = 0.5*((np.max(x))-(np.min(x)))\n    X_train.loc[segment, 'meanvar'] = x.std()/x.mean()\n    X_train.loc[segment , 'medianAbsDev'] = np.median(abs(x-np.median(x)))\n    X_train.loc[segment , 'percentAmp'] = np.max(np.abs(x-np.median(x)))/np.median(x)\n    X_train.loc[segment, 'andersonDarling'] = 1 / (1.0 + np.exp(-10 * (stats.anderson(x)[0] - 0.3)))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"from sklearn.model_selection import train_test_split\n\nXtrain, Xtest, ytrain, ytest = train_test_split(X_train,y_train, test_size=0.2)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_train.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"scaler = StandardScaler()\nscaler.fit(Xtrain)\nXtrain_scaled = scaler.transform(Xtrain)\nXtest_scaled = scaler.transform(Xtest)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"Xtrain_scaled.shape","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"lgb_train = lgb.Dataset(Xtrain_scaled, ytrain.values.flatten())\nlgb_eval = lgb.Dataset(Xtest_scaled, ytest, reference=lgb_train)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# specify your configurations as a dict\nparams = {\n    'boosting_type': 'gbdt',\n    'objective': 'regression',\n    'max_bin': 500,\n    'metric': {'l2', 'l1'},\n    'num_leaves': 200,\n    'learning_rate': 0.05,\n    'feature_fraction': 0.9,\n    'bagging_fraction': 0.8,\n    'bagging_freq': 5,\n    'verbose': 0\n}","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"start = time.time()\ngbm = lgb.train(params,\n                lgb_train,\n                num_boost_round=300,valid_sets=lgb_eval,\n                early_stopping_rounds=5)\nend = time.time()\nfit_time = (end - start)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"submission = pd.read_csv('../input/sample_submission.csv', index_col='seg_id')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test = pd.DataFrame(columns=X_train.columns, dtype=np.float64, index=submission.index)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test.index","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"for seg_id in X_test.index:\n    seg = pd.read_csv('../input/test/' + seg_id + '.csv')\n    \n    \n    x = seg['acoustic_data'].values\n    X_test.loc[seg_id, 'ave'] = x.mean()\n    X_test.loc[seg_id, 'std'] = x.std()\n    X_test.loc[seg_id, 'max'] = x.max()\n    X_test.loc[seg_id, 'min'] = x.min()\n    X_test.loc[seg_id, 'q95'] = np.quantile(x,0.95)\n    X_test.loc[seg_id, 'q99'] = np.quantile(x,0.99)\n    X_test.loc[seg_id, 'q05'] = np.quantile(x,0.05)\n    X_test.loc[seg_id, 'q01'] = np.quantile(x,0.01)\n    \n    X_test.loc[seg_id, 'abs_max'] = np.abs(x).max()\n    X_test.loc[seg_id, 'abs_mean'] = np.abs(x).mean()\n    X_test.loc[seg_id, 'abs_std'] = np.abs(x).std()\n    \n    X_test.loc[seg_id, 'iqr'] = np.subtract(*np.percentile(x, [75, 25]))\n    X_test.loc[seg_id, 'q999'] = np.quantile(x,0.999)\n    X_test.loc[seg_id, 'q001'] = np.quantile(x,0.001)\n    X_test.loc[seg_id, 'ave10'] = stats.trim_mean(x, 0.1)\n    \n    X_test.loc[seg_id, 'skew'] = stats.skew(x)\n    X_test.loc[seg_id, 'av_change_abs'] = np.mean(np.diff(x))\n    X_test.loc[seg_id, 'av_change_rate'] = np.mean(np.nonzero((np.diff(x) / x[:-1]))[0])\n\n# Extra calculated time series features\n    X_test.loc[seg_id, 'rcs'] = calRCS(x)\n    X_test.loc[seg_id, 'amplitude'] = 0.5*((np.max(x))-(np.min(x)))\n    X_test.loc[seg_id, 'meanvar'] = x.std()/x.mean()\n    X_test.loc[seg_id , 'medianAbsDev'] = np.median(abs(x-np.median(x)))\n    X_test.loc[seg_id , 'percentAmp'] = np.max(np.abs(x-np.median(x)))/np.median(x)\n    X_test.loc[seg_id, 'andersonDarling'] = 1 / (1.0 + np.exp(-10 * (stats.anderson(x)[0] - 0.3)))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test_scaled = scaler.transform(X_test)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_test_scaled","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"submission['time_to_failure'] = gbm.predict(X_test_scaled)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"submission.to_csv('submission.csv')","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}