A novel approach for breast cancer prediction using optimized ANN classifier based on big data environment

Area of Science:

• Oncology
• Bioinformatics
• Computer Science

Background:

• Breast cancer (BC) is a leading cause of mortality in women, necessitating improved prediction models.
• Current BC prediction models are often time-consuming and lack sufficient accuracy.
• Early and accurate prediction of breast cancer is crucial for effective treatment and patient outcomes.

Purpose of the Study:

• To develop an efficient and accurate Breast Cancer Prediction System (BCPS).
• To overcome the limitations of existing time-consuming and less accurate BC prediction models.
• To leverage Optimized Artificial Neural Network (OANN) for enhanced BC prediction.

Main Methods:

• Utilized Hadoop MapReduce for eliminating redundant information in big data (BD) storage.
• Applied data preprocessing techniques including replacing missing attributes (RMA) and normalization.
• Employed Modified Dragonfly algorithm (MDF) for feature selection and Gray Wolf Optimization (GWO) for OANN optimization.
• Classified selected features using the Optimized Artificial Neural Network (OANN).

Main Results:

• The proposed BCPS demonstrates superior performance compared to the prevailing Improved Weighted-Decision Tree (IWDT) model.
• Achieved higher precision, recall, accuracy, and ROC values in experimental outcomes.
• The OANN model, optimized with GWO, significantly enhances BC prediction accuracy.

Conclusions:

• The developed BCPS using OANN offers a more efficient and accurate approach to breast cancer prediction.
• The integration of big data processing, feature selection, and optimized neural networks shows significant promise.
• This system can aid in earlier and more reliable diagnosis of breast cancer, potentially improving patient survival rates.