Breast Cancer Prediction Based on Feature Extraction using Hybrid Methodologies

Abstract

The breast cancer prediction is essential for effective treatment and management of the disease. Using data mining techniques to develop predictive models can assist in identifying patients at high risk of developing breast cancer, allowing for early detection and treatment. Early detection has been shown to improve patient outcomes and survival rates. The proposed system for breast cancer prediction involves two main techniques: Linear Discriminant Analysis (LDA) based feature extraction and hyperparameter tuned LSTM-XGBoost based hybrid modelling. The LDA is used to extract the features from the input data that can be trained using a hybrid model such as LSTM and XGBoost. The hyperparameters of both models are optimized using cross-validation techniques to achieve high accuracy in breast cancer prediction. Overall, this proposed system has achieved an accuracy and efficiency of breast cancer prediction than existing.