Optimized Hyperparameters Machine Learning Model for Accurate and Efficient Breast Cancer Diagnosis System

Nowadays, breast cancer is the second most deadly type of cancer in humans, especially in females. Healthcare frequently uses machine learning (ML) to predict fatal diseases. We developed a model to leverage the Dataset on Wisconsin Diagnostic Breast Cancer (WDBC) because early detection is crucial for a good prognosis for breast cancer. On the flip side, the overall principle of this model is to evaluate seven popular ML classifiers: K-Nearest Neighbor, Naive Bayes, Adaboost, Logistic Regression, Decision Tree, Random Forest and Support Vector Machine. In order to mitigate the impact of traditional methods, our main strategy involved hyper parameter tweaking using the grid search method. There were a number of metrics that were enhanced as a result of this, including Torecision, recall, F1 score, area under the curve (AUC), and ROC curve. While the traditional method improved accuracy from 91.56% to 94.08%, the hyper parameter tuning model increased accuracy rate from 93.15% to 97.83%. The results showed that with an accuracy score of 97.83%, KNN was the most effective classifier. Finally, our experimental results demonstrate that KNN is compatible with the hyper tuning approach. Based on these findings, our research prediction strategy outperforms the standard method for accurately predicting breast cancer patients' prognosis.