Neural networks and LDA-based machine learning framework for the early detection of breast cancer

The International Agency for Research on Cancer (IARC) has released new statistics indicating that breast cancer has surpassed lung cancer and diagnosed as the most prevalent cancer in women. Breast cancer is one of the top causes of mortality for women. Early detection and supportive care can help to increase the survival rate. Lack of reliable prognostic models challenges doctors to design a treatment strategy that could prolong a patient's survival rate. In order to tackle this, scientists have developed a machine learning framework for early detection of breast cancer using the Wisconsin dataset. Thirty-two characteristics and 569 distinct cancer biopsy samples are included in the collection. There are 212 cases identified as malignant and 357 cases classed as benign in the dataset's class distribution. Four significant digits are used to record the feature values in the dataset; no attribute value is missing. The dataset was used to train a number of machine learning models, such as Random Forest, Support Vector Machine (SVM), Naive Bayes (NB), K Nearest Neighbors (K-NN), and Neural Networks. Linear discriminant analysis (LDA) and principal component analysis (PCA) were also employed, resulting in the discovery that neural networks with LDA outperform other methods. The Neural Network with LDA model's F1 score, sensitivity, and balanced accuracy were 96.72%, 93.65%, and 96.83%, respectively.