How tulsip26-hub/loan-prediction-model_classification works
This repository is a loan-approval classification project. Its documented workflow takes the loan dataset through cleaning, exploratory analysis, feature engineering, preprocessing and train/test splitting, then trains and evaluates classifiers, compares them, and identifies a best model. The README also describes feature-importance analysis. Source excerpts are empty, so the stage links reflect README documentation rather than verified implementation calls; the tree supports the listed module paths but not finer internal wiring.
Main entry points: model_build.py Train-test split · loan_data.csv Loan dataset · data_ingestion.py Data ingestion · data_preprocessing.py Data cleaning
Components 16
Data analysis
- Loan dataset
data/loan_data.csv - Data ingestion
src/data_ingestion.py - Data cleaning
src/data_preprocessing.py - Exploratory analysis
src
Feature preparation
- Feature engineering
src/data_preprocessing.py - Data preprocessing
src/data_preprocessing.py - Train-test split
src/model_build.py
Modeling and assessment
- Logistic regression
src/model_build.py - Decision tree
src/model_build.py - Random forest
src/model_build.py - Model evaluation
src/model_build.py - Model comparison
src/model_build.py - Best model
src/model_build.py - Feature importance
src/model_build.py
Other components
- Project user
- Approval prediction
Connections 17
- Project user → Loan dataset: provides data
- Loan dataset → Data ingestion: loads
- Data ingestion → Data cleaning: supplies data
- Data cleaning → Exploratory analysis: analyzes
- Exploratory analysis → Feature engineering: informs
- Feature engineering → Data preprocessing: creates features
- Data preprocessing → Train-test split: prepares data
- Train-test split → Logistic regression: trains
- Train-test split → Decision tree: trains
- Train-test split → Random forest: trains
- Logistic regression → Model evaluation: evaluated by
- Decision tree → Model evaluation: evaluated by
- Random forest → Model evaluation: evaluated by
- Model evaluation → Model comparison: scores models
- Model comparison → Best model: selects
- Best model → Feature importance: analyzed for
- Best model → Approval prediction: predicts approval