{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.14","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":81933,"databundleVersionId":9643020,"sourceType":"competition"}],"dockerImageVersionId":30786,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"#
🌐 AutoML H2O
\n\n\n![](https://i.postimg.cc/Jz5DS0dq/pexels-james-frid-81279-9823161.jpg)\n\n# ❗Understanding the Competition\n
\n

Aim of competition

\n

This competition aimed to predict the severity of Problematic Internet Use (PIU) in children and adolescents based on their physical activity and fitness data.

\n\n

Key Points:

\n \n
\n\n# πŸŽ’ Import Libraries","metadata":{}},{"cell_type":"code","source":"import numpy as np \nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nfrom sklearn.model_selection import StratifiedKFold\nfrom scipy.optimize import minimize\nimport os\nfrom tqdm import tqdm\nfrom IPython.display import clear_output\nfrom concurrent.futures import ThreadPoolExecutor\nfrom sklearn.preprocessing import StandardScaler, LabelEncoder\nimport h2o\nfrom h2o.automl import H2OAutoML\nfrom sklearn.base import clone\nfrom sklearn.metrics import *\nfrom colorama import Fore, Style\n\nSEED = 42\nn_splits = 5","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2024-10-30T18:01:41.584197Z","iopub.execute_input":"2024-10-30T18:01:41.584664Z","iopub.status.idle":"2024-10-30T18:01:41.592792Z","shell.execute_reply.started":"2024-10-30T18:01:41.584622Z","shell.execute_reply":"2024-10-30T18:01:41.591528Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Python Modules","metadata":{}},{"cell_type":"markdown","source":"
\n

Imported Modules

\n \n\n

Global Variables

\n \n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ“€ Load CSV Files","metadata":{}},{"cell_type":"code","source":"train_df = pd.read_csv('/kaggle/input/child-mind-institute-problematic-internet-use/train.csv')\nsample = pd.read_csv('/kaggle/input/child-mind-institute-problematic-internet-use/sample_submission.csv')\n\nTARGET_COLS = [\n \"PCIAT-Season\",\n \"PCIAT-PCIAT_01\",\n \"PCIAT-PCIAT_02\",\n \"PCIAT-PCIAT_03\",\n \"PCIAT-PCIAT_04\",\n \"PCIAT-PCIAT_05\",\n \"PCIAT-PCIAT_06\",\n \"PCIAT-PCIAT_07\",\n \"PCIAT-PCIAT_08\",\n \"PCIAT-PCIAT_09\",\n \"PCIAT-PCIAT_10\",\n \"PCIAT-PCIAT_11\",\n \"PCIAT-PCIAT_12\",\n \"PCIAT-PCIAT_13\",\n \"PCIAT-PCIAT_14\",\n \"PCIAT-PCIAT_15\",\n \"PCIAT-PCIAT_16\", \n \"PCIAT-PCIAT_17\",\n \"PCIAT-PCIAT_18\",\n \"PCIAT-PCIAT_19\",\n \"PCIAT-PCIAT_20\",\n \"PCIAT-PCIAT_Total\"\n]\n\ntrain_df = train_df.drop(TARGET_COLS,axis=1)\n\ntest_df = pd.read_csv('/kaggle/input/child-mind-institute-problematic-internet-use/test.csv')\nids = test_df['id']\n\nSEASON_COLS = [\n \"Basic_Demos-Enroll_Season\", \"CGAS-Season\", \"Physical-Season\",\n \"Fitness_Endurance-Season\", \"FGC-Season\", \"BIA-Season\",\n \"PAQ_A-Season\", \"PAQ_C-Season\", \"SDS-Season\", \"PreInt_EduHx-Season\"\n]\n\ntrain_df = train_df.drop(SEASON_COLS,axis=1)\ntest_df = test_df.drop(SEASON_COLS,axis=1)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:01:41.638208Z","iopub.execute_input":"2024-10-30T18:01:41.638750Z","iopub.status.idle":"2024-10-30T18:01:41.709540Z","shell.execute_reply.started":"2024-10-30T18:01:41.638698Z","shell.execute_reply":"2024-10-30T18:01:41.707882Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Load CSV Files","metadata":{}},{"cell_type":"markdown","source":"
\n

Data Loading

\n \n\n

Target Columns

\n \n\n

Dropping Target Columns

\n\n\n

Loading Test Data and Extracting IDs

\n\n

Season Columns

\n \n\n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ“€ Load Parquet Files","metadata":{}},{"cell_type":"code","source":"def process_file(filename, dirname):\n df = pd.read_parquet(os.path.join(dirname, filename, 'part-0.parquet'))\n df.drop('step', axis=1, inplace=True)\n return df.describe().values.reshape(-1), filename.split('=')[1]\n\ndef load_time_series(dirname) -> pd.DataFrame:\n ids = os.listdir(dirname)\n \n with ThreadPoolExecutor() as executor:\n results = list(tqdm(executor.map(lambda fname: process_file(fname, dirname), ids), total=len(ids)))\n \n stats, indexes = zip(*results)\n \n df = pd.DataFrame(stats, columns=[f\"Stat_{i}\" for i in range(len(stats[0]))])\n df['id'] = indexes\n \n return df\n\ntrain_ts = load_time_series(\"/kaggle/input/child-mind-institute-problematic-internet-use/series_train.parquet\")\ntest_ts = load_time_series(\"/kaggle/input/child-mind-institute-problematic-internet-use/series_test.parquet\")\ntime_series_cols = train_ts.columns.tolist()\ntime_series_cols.remove(\"id\")","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:01:41.711898Z","iopub.execute_input":"2024-10-30T18:01:41.712443Z","iopub.status.idle":"2024-10-30T18:03:35.477564Z","shell.execute_reply.started":"2024-10-30T18:01:41.712387Z","shell.execute_reply":"2024-10-30T18:03:35.476196Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Load Parquet Files","metadata":{}},{"cell_type":"markdown","source":"
\n\n

Process Function (process_file)

\n\n \n\n

Load Time Series (load_time_series)

\n\n \n\n

Loading Training and Test Time Series

\n \n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ—’οΈ Merge CSV & Parquet","metadata":{}},{"cell_type":"code","source":"train = pd.merge(train_df, train_ts, how=\"left\", on='id')\ntest = pd.merge(test_df, test_ts, how=\"left\", on='id')\n\ntrain = train.drop('id', axis=1)\ntest = test.drop('id', axis=1)\ntrain = train.dropna(subset=['sii'])","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.479651Z","iopub.execute_input":"2024-10-30T18:03:35.480186Z","iopub.status.idle":"2024-10-30T18:03:35.509557Z","shell.execute_reply.started":"2024-10-30T18:03:35.480138Z","shell.execute_reply":"2024-10-30T18:03:35.508238Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Merge CSV & Parquet\n\n
\n

Merging Tabular and Time Series Data

\n \n\n

Dropping Unnecessary Column

\n \n\n

Handling Missing Values

\n \n
","metadata":{}},{"cell_type":"markdown","source":"# ✨ Preprocessing","metadata":{}},{"cell_type":"code","source":"def preprocess_data(df,train_data=False):\n # Handle numerical columns\n num_cols = df.select_dtypes(include=np.number).columns\n df[num_cols] = df[num_cols].fillna(df[num_cols].median())\n \n return df\n\ntrain = preprocess_data(train)\ntest = preprocess_data(test)\n\ntrain = pd.DataFrame(train)\ntest = pd.DataFrame(test)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.511308Z","iopub.execute_input":"2024-10-30T18:03:35.511708Z","iopub.status.idle":"2024-10-30T18:03:35.696445Z","shell.execute_reply.started":"2024-10-30T18:03:35.511667Z","shell.execute_reply":"2024-10-30T18:03:35.695244Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Preprocessing\n
\n

Data Preprocessing and Feature Engineering

\n\n

This code snippet defines a function: preprocess_data to handle features, respectively.

\n\n

Preprocessing Data

\n

The preprocess_data function performs the following steps:

\n \n\n

Finally, the preprocessed training and testing DataFrames are converted back to Pandas DataFrames.

\n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ”Ž View Data","metadata":{}},{"cell_type":"code","source":"train.head()","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.709137Z","iopub.execute_input":"2024-10-30T18:03:35.709661Z","iopub.status.idle":"2024-10-30T18:03:35.760818Z","shell.execute_reply.started":"2024-10-30T18:03:35.709605Z","shell.execute_reply":"2024-10-30T18:03:35.759445Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test.head()","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.762610Z","iopub.execute_input":"2024-10-30T18:03:35.763042Z","iopub.status.idle":"2024-10-30T18:03:35.802397Z","shell.execute_reply.started":"2024-10-30T18:03:35.762986Z","shell.execute_reply":"2024-10-30T18:03:35.800888Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: View Data\n\n
\n

Displaying First Few Rows of DataFrames

\n \n
","metadata":{}},{"cell_type":"markdown","source":"# βš–οΈ Quadratic Weighted Kappa","metadata":{}},{"cell_type":"code","source":"def quadratic_weighted_kappa(y_true, y_pred):\n return cohen_kappa_score(y_true, y_pred, weights='quadratic')\n\ndef threshold_Rounder(oof_non_rounded, thresholds):\n return np.where(oof_non_rounded < thresholds[0], 0,\n np.where(oof_non_rounded < thresholds[1], 1,\n np.where(oof_non_rounded < thresholds[2], 2, 3)))\n\ndef evaluate_predictions(thresholds, y_true, oof_non_rounded):\n rounded_p = threshold_Rounder(oof_non_rounded, thresholds)\n return -quadratic_weighted_kappa(y_true, rounded_p)\n\ndef TrainML(model_class, test_data, train_data):\n X = train_data.drop(['sii'], axis=1)\n y = train_data['sii']\n\n SKF = StratifiedKFold(n_splits=n_splits, shuffle=True, random_state=SEED)\n \n train_S = []\n test_S = []\n \n oof_non_rounded = np.zeros(len(y), dtype=float) \n oof_rounded = np.zeros(len(y), dtype=int) \n test_preds = np.zeros((len(test_data), n_splits))\n\n for fold, (train_idx, test_idx) in enumerate(tqdm(SKF.split(X, y), desc=\"Training Folds\", total=n_splits)):\n X_train, X_val = X.iloc[train_idx], X.iloc[test_idx]\n y_train, y_val = y.iloc[train_idx], y.iloc[test_idx]\n\n X_train = h2o.H2OFrame(X_train)\n y_train_pred = best_model.predict(X_train)\n y_train_pred = y_train_pred.as_data_frame()\n y_train_pred = y_train_pred.values\n y_train_pred = y_train_pred.reshape(-1)\n \n X_val = h2o.H2OFrame(X_val)\n y_val_pred = best_model.predict(X_val)\n y_val_pred = y_val_pred.as_data_frame()\n y_val_pred = y_val_pred.values\n y_val_pred = y_val_pred.reshape(-1)\n\n oof_non_rounded[test_idx] = y_val_pred\n y_val_pred_rounded = y_val_pred.round(0).astype(int)\n oof_rounded[test_idx] = y_val_pred_rounded\n\n train_kappa = quadratic_weighted_kappa(y_train, y_train_pred.round(0).astype(int))\n val_kappa = quadratic_weighted_kappa(y_val, y_val_pred_rounded)\n\n train_S.append(train_kappa)\n test_S.append(val_kappa)\n \n test_data_copy = h2o.H2OFrame(test_data)\n test_data_copy = best_model.predict(test_data_copy)\n test_data_copy = test_data_copy.as_data_frame()\n test_data_copy = test_data_copy.values\n test_data_copy = test_data_copy.reshape(-1)\n test_preds[:, fold] = test_data_copy \n \n print(f\"Fold {fold+1} - Train QWK: {train_kappa:.4f}, Validation QWK: {val_kappa:.4f}\")\n clear_output(wait=True)\n\n print(f\"Mean Train QWK --> {np.mean(train_S):.4f}\")\n print(f\"Mean Validation QWK ---> {np.mean(test_S):.4f}\")\n\n KappaOPtimizer = minimize(evaluate_predictions, x0=[0.5, 1.5, 2.5], args=(y, oof_non_rounded), method='Nelder-Mead') # Nelder-Mead | # Powell\n assert KappaOPtimizer.success, \"Optimization did not converge.\"\n \n oof_tuned = threshold_Rounder(oof_non_rounded, KappaOPtimizer.x)\n tKappa = quadratic_weighted_kappa(y, oof_tuned)\n\n print(f\"----> || Optimized QWK SCORE :: {Fore.CYAN}{Style.BRIGHT} {tKappa:.3f}{Style.RESET_ALL}\")\n\n tpm = test_preds.mean(axis=1)\n tpTuned = threshold_Rounder(tpm, KappaOPtimizer.x)\n \n submission = pd.DataFrame({\n 'id': sample['id'],\n 'sii': tpTuned\n })\n\n return submission,KappaOPtimizer","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.804823Z","iopub.execute_input":"2024-10-30T18:03:35.805307Z","iopub.status.idle":"2024-10-30T18:03:35.829373Z","shell.execute_reply.started":"2024-10-30T18:03:35.805255Z","shell.execute_reply":"2024-10-30T18:03:35.826169Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Quadratic Weighted Kappa\n\n
\n

Quadratic Weighted Kappa and Training Function

\n\n

This section defines two functions: quadratic_weighted_kappa and TrainML.

\n\n

quadratic_weighted_kappa

\n \n\n

TrainML

\n

This function is responsible for training a machine learning model and evaluating its performance. It takes the model class (model_class), testing data (test_data), and training data (train_data) as input.

\n\n \n
","metadata":{}},{"cell_type":"markdown","source":"# 🌐 AutoML H2O","metadata":{}},{"cell_type":"code","source":"h2o.init()\ntrain_data = h2o.H2OFrame(train)\n\naml = H2OAutoML(max_runtime_secs=5400,seed=5)\naml.train(y='sii', training_frame=train_data)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:03:35.832495Z","iopub.execute_input":"2024-10-30T18:03:35.833186Z","iopub.status.idle":"2024-10-30T18:04:08.577781Z","shell.execute_reply.started":"2024-10-30T18:03:35.833122Z","shell.execute_reply":"2024-10-30T18:04:08.576438Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: AutoML H2O\n\n
\n

Initializing H2O and Training an AutoML Model

\n\n

This code snippet initializes the H2O environment and trains an AutoML model on the training data.

\n\n \n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ† Leaderboard","metadata":{}},{"cell_type":"code","source":"leaderboard = aml.leaderboard","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:04:08.579420Z","iopub.execute_input":"2024-10-30T18:04:08.579799Z","iopub.status.idle":"2024-10-30T18:04:08.586287Z","shell.execute_reply.started":"2024-10-30T18:04:08.579759Z","shell.execute_reply":"2024-10-30T18:04:08.584751Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Leaderboard\n\n
\n

Retrieving the AutoML Leaderboard

\n\n \n\n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ₯‡ Best Models","metadata":{}},{"cell_type":"code","source":"best_model = aml.leader\nSubmission,KappaOPtimizer = TrainML(best_model,test,train)\nprint(KappaOPtimizer.x)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:04:08.590550Z","iopub.execute_input":"2024-10-30T18:04:08.591117Z","iopub.status.idle":"2024-10-30T18:04:18.731388Z","shell.execute_reply.started":"2024-10-30T18:04:08.591056Z","shell.execute_reply":"2024-10-30T18:04:18.730067Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Best Models\n\n
\n

Selecting the Best Model and Making Predictions

\n\n \n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ“ Prepare Submission CSV","metadata":{}},{"cell_type":"code","source":"Submission.to_csv('submission.csv', index=False)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:04:18.732898Z","iopub.execute_input":"2024-10-30T18:04:18.733324Z","iopub.status.idle":"2024-10-30T18:04:18.741417Z","shell.execute_reply.started":"2024-10-30T18:04:18.733273Z","shell.execute_reply":"2024-10-30T18:04:18.740054Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: Prepare Submission CSV\n\n
\n

Creating the Submission File

\n\n \n
","metadata":{}},{"cell_type":"markdown","source":"# πŸ“° View Results","metadata":{}},{"cell_type":"code","source":"print(Submission['sii'].value_counts())\nSubmission.head(20)","metadata":{"execution":{"iopub.status.busy":"2024-10-30T18:04:18.743371Z","iopub.execute_input":"2024-10-30T18:04:18.743768Z","iopub.status.idle":"2024-10-30T18:04:18.769794Z","shell.execute_reply.started":"2024-10-30T18:04:18.743725Z","shell.execute_reply":"2024-10-30T18:04:18.768440Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# πŸ“‘ Explaining: View Results\n\n
\n

Analyzing and Visualizing the Submission

\n \n
","metadata":{}}]}