Neural Network for Breast Cancer Prediction
International Journal of Advanced Trends in Computer Science and Engineering · 2020
The aim of this study was to establish a method for breast cancer prediction, mainly based on blood test results.The study utilized data collected from 166 participants, half of whom were diagnosed with breast cancer.The data consist of information like the body mass index (BMI), age, and results of seven blood tests that encircle homeostatic model assessment (HOMA) and analyses of glucose, insulin, leptin, adiponectin, resistin, and monocyte chemotactic protein 1 (MCP-1).These seven blood tests are considered as indicators of breast cancer.However, this study was intended to identify the particular parameters of these that are most indicative of breast cancer.Two experiments were conducted by using the feed-forward, back-propagation artificial neural network.The first experiment considered all the foregoing nine parameters while the second experiment was run on only seven parameters because a decision tree for feature selection identified seven of the original nine parameters as strong predictors of breast cancer.Values of the prediction accuracy were 93.1% in the first experiment with nine predictors and 94% in the second experiment with seven predictors.Moreover, specificity and sensitivity were calculated to assess the performance of the proposed method with the two sets of predictors.The indications are promising.As such, this study has the contribution of providing rapid, non-invasive means of breast cancer prediction for early treatment.