An Enhanced Modular-Based Neural Network Framework for Effective Medical Diagnosis

Abstract

Artificial Neural Networks (ANNs) are a type of machine learning algorithms that are used to solve problems such as medical diagnosis. In recent times, the amount of data that is generated daily is on the increase and the level of the complexity of problems is troubling. ANN algorithms are commonly used to overcome these challenges are further faced with the problem of having fixed data attributes as a dataset for its input layer, the complexity of having heterogeneous datasets instead of homogeneous datasets,and having a single objective output layer instead of a multi-objective output layer that could enable the diagnosis of multiple diseases. This researchproposes an enhanced modular-based Neural Network algorithm that utilizes heterogeneous datasets drawn from multiple sources, decomposed and clustered into independent units, and then trained by ANNs selected according to their learning paradigms -supervised, unsupervised, and reinforcement learning, to provide an effective, efficient and timely medical diagnosis, especially in developing countries where modern facilities are lacking with much dependence on manual methods.Thus an integrated system with multiple ANN techniques modelled into a single unit is developed. The results show that the proposed approach has been significantly successful indealing with the aforesaid problem compared to other methods with a training accuracy of 0.905, Sensitivity of 0.917, and specificity of 0.923.