An optimized K-Nearest neighbor based breast cancer detection

In this research, a grid search is employed to find the optimal hyper-parameter and an optimized K-Nearest Neighbor (KNN) based breast cancer detection model is proposed. The grid search is employed to find the best value of K that could produce better breast cancer detection accuracy. Moreover, this study explored the effect of hyper-parameter tuning on the performance of KNN for breast cancer detection. The findings of this research reveals that hyper-parameter tuning has a significant effect on the performance of the KNN model. The effect of hyper-parameter tuning on the performance of KNN algorithm is experimentally tested using Wisconsin breast cancer dataset collected from kaggle data repository. Finally, we have compared the performance of the KNN with the tuned hyper-parameter and with default hyper-parameter. The result analysis on the performance of the model on breast cancer detection using the testing set reveals that the accuracy of the proposed optimized model is 94.35% and the performance of the KNN with the default hyper-parameter is 90.10%.