Machine Learning for Bone Tumor Diagnostics

This project involved developing and evaluating machine learning models to classify bone tumor patient status into three categories:

No Evidence of Disease (NED)

Alive with Disease (AWD)

Deceased (D)

The models were trained and tested on a dataset from the Memorial Sloan Kettering Cancer Center with 500 samples and 9 features.

The following machine learning classification algorithms were implemented and evaluated:

Gaussian Naive Bayes

Decision Tree

Random Forest

Logistic Regression

Linear SVC

K-Nearest Neighbors (Euclidean and Manhattan distances)

The models were thoroughly evaluated on classification performance metrics:

Accuracy

Precision

Recall

F1 Score

Execution Time

Among all the models, Logistic Regression achieved the highest accuracy of 81% in classifying bone tumor patient status.

Linear SVC also performed well in minimizing false negatives. In contrast, Gaussian Naive Bayes and KNN models showed limitations in distinguishing between patient statuses.

This repository contains:

Jupyter notebooks for data preprocessing, EDA, and model implementation

Model evaluation results

The dataset is from Kaggle: https://www.kaggle.com/datasets/antimoni/bone-tumor