Analysing Breast Cancer Classification Using Explainable Artificial Intelligence

Worldwide, breast cancer is becoming the most serious illness that affects women. It is believed that early diagnosis and treatment of breast cancer can increase survival rates and decrease the need for surgery. Machine Learning model is very reliant on features for their proper training. However, understanding how a prediction is being affected by specific features is very important for a model’s interpretation. Understanding what features support a prediction is important as it provides some transparency to the inner workings of the model. Gaussian Naive Bayes (GNB), Decision Tree (DT), Random Forest (RF), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM) are used to classify the breast cancer data, and accuracy, sensitivity, specificity, false-positive rate, precision, F<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf>-score, and Geometric-Mean (GM) are used for the performance assessment. Furthermore, Multi-Criteria Decision Making (MCDM) is used to evaluate overall performance assessment based on the aforementioned performance measures and DT is found to be best among all the classifiers. Finally, an Explainable AI model namely LIME is used to interpret the predicted outcomes and impact of the different features of the data on the model’s prediction.