https://www.kaggle.com/ashishpatel26/home-credit-default-risk-001
Assessing approval risk for home credit applications
Title /Abstract / Block Diagram - Nishad Data you plan to use/ ER Diagram - Vishal Machine Learning Algo / Metrics to measure - Naimesh Description of pipelines -Nishad List team member - VIshal
Indication - Nishad EDA Feature Engineering - Vishal Data Formatting / Feature Engineering Primary - EDA (Secondary) - Nishad Model selection / Hyper Parameter tuning / Feature Engineering - Naimesh
Feature engineering -
Bureau and bureau balance - Vishal
Credit card , Installments, previous applications - Nishad
Application train , POS_CASH_Balance , Previous applications - Naimesh
Phase - 1 submission work distribution
Downloaded the data correctly - Done (Kaggle Part)
Good EDA - In progress - Vishal
Additional feature engineering. Identification and development of additional features - In progress - Summary statistics
Use of Pipeline - Done
Data prep - Done Appropriate us of train, validate and test splits. Filled in missing data, aggregated data appropriately (e.g., aggregated by member ID/year to generate predictions for a given individual in a given year).
Experimental results table - Friday
Statistical significance test - Friday Baseline vs. challenger. Best "new" model versus prior best
Input Features/Feature Selection - Naimesh / Nishad Evaluation and selection of best features for the model.
Discussion/analysis of results.
Project/team participation
ppt - Nishad - Friday
Consolidated notebook - Naimesh