Automated Prognostic Modelling of Alzheimer’s Disease Prediction based on Machine Learning over Brain Networks

Abstract

Medical Institutions rely on implementing E-Systems for tracking, monitoring and simplifying the process of diagnosing the life threatening diseases. Many advanced electronic health monitoring systems have become sophisticated with immense development in computing and communication technologies. Machine Learning is another technique that has a superior command over image processing, diagnosis and disease predictions. This research article emphasizes on the technique to provide an intelligent and novel technique to predict the progress of Alzheimer’s disease through a decision support system. The original raw images from the MRI devices are converted into a functional matrix which illustrates the 90 credible brain regions and thus their activities. Once the images are tracked down to their images and operational centres, the association between various regions is represented as a correlation matrix of the functional areas. This correlation process is computed for every pair of brain regions from the overall images. An automated encoder network is built to identify and classify the regions affected in close relation to the Alzheimer’s disease. The experimental set up evaluates the proposed mechanism against the conventional models of prediction and the results reveal that the proposed scheme is faster and reliable than the conventional strategies. The correlations between different regions of the brain achieved with an improved accuracy to detect the Alzheimer’s disease. This proposal has shown significant increase of 25% betterments when compared to Support Vector Machine (SVM). This benefit has improvised the standards of electronic support systems in medical industries and prediction of Alzheimer’s Disease at early stages facilitates the diagnosis process pre-emptively. The onset of Alzheimer’s Disease should be limited to primary stages provided the method is applicable in real life applications.