README.md

This is the repo of Stat_Collective for the workings in Datastorm 2.0

Approach

• Load Dataset directly as csv into panda dataframe
• Train set, validation set and test set is loaded seperately.

1. Data Cleaning and EDA

• Duplicate dataset for undersampled (Skip this for first fit)
• Check for Missing Data
• Check for ordinal data masked us numerical data
• Plots and charts for Data

2. Data Preprocessing

• Dummy variables
• Impute variables?
• Feature Selection
• Scaling

3. Model Building

3.1 Logistic Regression

3.2 Random Forest Classification

3.3 K-Nearest Neighbours

3.4 Gradient Boosting

3.5 XG Boost

3.6 Support Vector Machine

3.7 Neural Network

4. Model Evaluation (Hyper Parameter Tuning)

• F1 score
• confusion matrix

5. Model Testing

• F1 score
• Confusion matrix