How himanshu-97-cloud/cancer-risk-prediction-ml works
This repository is a notebook-based machine-learning project for predicting cancer risk from patient health and lifestyle data. Its documented workflow runs from dataset exploration and cleaning through feature scaling, model comparison, imbalance handling, tuning, and final model selection. The README identifies RandomOverSampler, GridSearchCV, and AdaBoost as the best pipeline and reports final metrics. The notebook source was not sampled, so stage-level links follow the documented workflow rather than verified code calls.
Main entry points: Cancer_Risk_Prediction.ipynb Imbalance methods · cancer_dataset.csv Cancer dataset
Components 14
Data preparation
- Cancer dataset
data/cancer_dataset.csv - Exploratory analysis
notebook/Cancer_Risk_Prediction.ipynb - Data cleaning
notebook/Cancer_Risk_Prediction.ipynb - Feature scaling
notebook/Cancer_Risk_Prediction.ipynb
Model experiments
- Model comparison
notebook/Cancer_Risk_Prediction.ipynb - 12 classifiers
notebook/Cancer_Risk_Prediction.ipynb - Imbalance methods
notebook/Cancer_Risk_Prediction.ipynb - SMOTE
notebook/Cancer_Risk_Prediction.ipynb - RandomOverSampler
notebook/Cancer_Risk_Prediction.ipynb - ADASYN
notebook/Cancer_Risk_Prediction.ipynb
Selection and results
- GridSearchCV tuning
notebook/Cancer_Risk_Prediction.ipynb - AdaBoost selection
notebook/Cancer_Risk_Prediction.ipynb - Final metrics
notebook/Cancer_Risk_Prediction.ipynb
Other components
- Analyst
Connections 13
- Analyst → Exploratory analysis: runs notebook
- Cancer dataset → Exploratory analysis: provides data
- Exploratory analysis → Data cleaning: informs
- Data cleaning → Feature scaling: prepares data
- Feature scaling → Model comparison: feeds
- Model comparison → 12 classifiers: evaluates
- Model comparison → Imbalance methods: compares
- Imbalance methods → SMOTE: includes
- Imbalance methods → RandomOverSampler: includes
- Imbalance methods → ADASYN: includes
- RandomOverSampler → GridSearchCV tuning: best technique
- GridSearchCV tuning → AdaBoost selection: selects
- AdaBoost selection → Final metrics: yields