Accuracy Comparison of Data Mining Algorithms Used in the Diagnosis of Breast Cancer: A Scoping Review Study

Abstract

Introduction: Breast cancer recognized as one of the widespread types of invasive cancer. Early diagnosis of breast cancer is crucial in treating it. The concept of data mining refers to the process of discovering and identifying information in large datasets. Data mining uses a set of techniques and algorithms that can be used for the early detection of breast cancer. The present study gives comparisons between the performances of various algorithms used to diagnose breast cancer. Method: This scoping review was led by the framework of the JBI methodology rules. The search was conducted in some relevant electronic databases, and PICO based extracted data were analyzed with Excel software. Results: The most commonly used algorithms were SVM (8 cases), j48 and Naive Bayes (7 cases), and MLP (6 cases), and 34 cases were only used in one study. The accuracy rate obtained with FSRAIRS2 (100%) is the highest among the other reported algorithms by other researchers. Moreover, Canopy was the less accurate algorithm (accuracy= 65%). Conclusion: Any use of a data mining and knowledge discovery method on a data set requires some discussion on the accuracy of the extracted model on some test data. In this study, we have investigated 48 common algorithms on one of the most crucial areas in medicine. Using algorithms that have high accuracy, automated, and semi-automated tools can be designed and used by professionals for the timely detection of breast cancer.