Breast cancer detection

Problem 1 )Data Consists of Outlier in several Columns Problem 2) Data consists of Left sweeknews in several Columns Problem 3) Data consists of Zero or Missing values in several Columns Problem 4) Data is unbalanced in terms of Cancer(“0”) counts 234 and Not Cancer (“1”) counts 42 [ Ratio 5:1]

Machine Learning Prediction Results

I have applied 6 different Machine learning model with Unbalanced and Balanced data

Machine learning models

RandomForest Naive Bayes SVM KNN Logistic Regression Xgboost

Data Balance Techniques

Downsampling Techniques Near Miss

Upsampling Techniques 1) SMOTETomek 2) RandomOverSampler

In last slide I have shown the result of all machine learning models with unbalanced and balanced data

Red colour represent Highest accuracy on the basis of F1 score with existing models and techniques

Blue colour represent Second Highest accuracy on the basis of F1 score with existing models and techniques