Santander customer transaction prediction
Binary Classification Problem under Supervised Machine Learning using AUC-ROC Score for model evaluation.Training data consists of 2,00,000 observations and
Binary Classification Problem under Supervised Machine Learning using AUC-ROC Score for model evaluation.Training data consists of 2,00,000 observations and 200 numerical anonymous predictor features, target variable and transaction_ID with no missing values.Exploratory Data Analysis include visualization of descriptive features related to numerical features and outlier Analysis.Feature Engineering includes feature addition and Principal Component Analysis.Finally, using Light GBM with K-Fold Stratification reaching AUC-ROC Score of 0.91.