How mihirj2002/airquality_index_predictionmodel works
This repository is an air-quality regression project that predicts Relative Humidity (RH) from sensor measurements. Its documented workflow runs from dataset ingestion through data understanding, outlier handling, feature preparation, model training, and evaluation. The sampled Python modules confirm Excel loading, winsorization, an RH feature/target split, train/test scaling, and K-nearest-neighbor regression with an R² score. The README documents broader analysis and model-comparison capabilities; their detailed implementation and wiring cannot be confirmed from the sampled source.
Main entry points: Data_Preprocessing.py Feature scaling · Model_building.py KNN regressor · Data_Ingestion.py Data ingestion · model.ipynb Model comparison
Components 14
Data analysis
- Air quality data
data/AirQualityUCI.xlsx - Data ingestion
src/Data_Ingestion.py - Exploratory analysis
experiment/model.ipynb - Quality assessment
experiment/model.ipynb
Data preparation
- Outlier treatment
src/Data_Preprocessing.py - Feature preparation
src/Data_Preprocessing.py - Train test split
src/Data_Preprocessing.py - Feature scaling
src/Data_Preprocessing.py
Modeling
- Regression models
experiment/model.ipynb - KNN regressor
src/Model_building.py
Evaluation
- Model comparison
experiment/model.ipynb - Prediction diagnostics
experiment/model.ipynb - R² score
src/Model_building.py
Other components
- Analyst
Connections 14
- Analyst → Data ingestion: starts workflow
- Air quality data → Data ingestion: reads
- Data ingestion → Exploratory analysis: provides data
- Exploratory analysis → Quality assessment: assesses
- Quality assessment → Outlier treatment: examines data
- Outlier treatment → Feature preparation: prepares data
- Feature preparation → Train test split: separates RH
- Train test split → Feature scaling: splits data
- Feature scaling → Regression models: provides features
- Feature scaling → KNN regressor: provides arrays
- KNN regressor → R² score: computes R²
- Regression models → Model comparison: compares models
- KNN regressor → Model comparison: contributes score
- Model comparison → Prediction diagnostics: analyzes predictions