- ML based approach to predcit the loan status
- One-hot encoding
- Executed Baseline model (LGBM, Random Forest, Logistic regession)
- Simple ANN model comparison
- Feature space managed using target encoding for the categorical variables
- ANOVA testing to target variable to perfomr target encoding saving feature space
- data normalisation, correlation
- comparing models (linear regression, XGBoost, Random Forest Regressor, LGBM regressor)
- Explored PCA
- Algorithms explored and compared are Random Forest, LGBM
- Comapred Grid search and Random search
- Detailed exploration of the data set and Model
- Outlier detection and its visualisation
- Feature exploration
- Manual hyperparameter tuning to understand the effects of the parameters on the model
- Explored Random Forest; PCA; LGBM.
First ever competition on Kaggle, an to competitions, a time series analysis for
- Time Analysis and Regression analysis
- AR/MA time model
- ARIMA / SARIMA model wiht various orders
- Accuracy with Mean Absolute Percentage Error(MAPE)
- Robust Scaling (Outliers)
- Ridge Regression (Multicollinearity)