Map-Centric Belief Rule Based System to Assess Dyslexia with Uncertainty

Abstract

Dyslexia is a serious problem on students in all respects. During this digitalization era, competitions are increasing day by day with enormous stress. Automated tool will be helpful to predict the dyslexia symptoms. Machine learning techniques with artificial intelligences are significant to develop the framework. However, excessive datasets on current medical diagnosis is a big challenge to design the system with numerous uncertainties. Map-Centric Belief Rule Based system (MapBRB) is an integrated framework that handles large datasets as well as predicts the symptoms of dyslexia at tertiary level students. Map-Centric phase clean the noises and irregularities from collected datasets. This process constructed by two pivotal parts as mapping and reducing. Mapping organizes the datasets and reducing part removes the irrelevant factors. Followed by Map-Centric, BRB operates the with inference engine (IE) and evidential reasoning(ER). During this experiment, uncertainty factors for dyslexic students are also checked and addressed. Few factors of dyslexia symptoms lead vague meaning and confusing. In this regards, BRB process the vague data with appropriate dimensions. Both qualitative and quantitative datasets have been addressed. Substantial comparisons have been performed among MapBRB and mapping less fuzzy system, mapping less BRB system and mapping less prediction system.