An Ensemble Classification Model for the Diagnosis of Breast Cancer Using Stacked Generalization

Breast cancer is one of the most common types of cancer whose incidence has increased dramatically in recent years. In order to diagnose this disease, many parameters must be taken into consideration and mistakes are possible due to human errors or environmental factors. For this reason, in recent decades, Artificial Intelligence has been used by medical practitioners to diagnose this disease.

In this applied-descriptive study, the diagnosis of breast cancer using stacked generalization was presented in the form of an ensemble model based on MLP neural network, ID3 decision tree, and support vector machine methods. To improve the performance of the ensemble classification model, a new approach called separator block was used. This block is responsible for identifying instances that cause errors in the classification model.

In order to evaluate the accuracy of the proposed method, the Wisconsin database for breast cancer was used. The experimental results showed the superiority of the proposed method over other similar methods. The accuracy of the classification model presented on the WBCD, WDBC, and WPBC datasets from the Wisconsin database was 99.54%, 99.58% and 99.84%, respectively.

Data mining algorithms can provide new and more cost-effective systems in the field of health and treatment that can diagnose breast cancer with high accuracy. In this study, modeling based on the stacked generalization technique was of high accuracy in the diagnosis of breast cancer.