3-Satisfiability Reverse Analysis Method with Hopfield Neural Network for Medical Data Set
DOI:
https://doi.org/10.4314/cajost.v2i1.37Keywords:
3-Satisfiability, Discrete Hopfield neural network, Medical Dataset, Data mining, Logic programmingAbstract
3-Satisfiability Reverse Analysis Method (3SATRA) incorporated with Hopfield neural network is a new approach for the early screening based on the medical dataset. 3SATRA will be proposed to extract the best logic rule that will represent the features of a medical dataset since the conventional data extraction techniques focus only on standalone neural network. The 3-Satisfiability Reverse Analysis Method will be integrated with logic programming as a data mining instrument. The proposed method is applied to Breast Cancer Coimbra and Statlog (Heart) dataset obtained from UCI machine learning repository. To pursue that, the results of the analysis will promote the early screening stage used for medical practitioners. The simulation will be executed using Dev C++ 5.11 as a tool for training, testing and validating the performances of the proposed method. The performance of the method was measured based on Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Sum of Squared Error (SSE), and Computational Time. The performance and accuracy of the results obtained have shown the effectiveness of 3SATRA in medical data mining.