Classification and Subgroup Analysis of Dyslexic Individuals

Abstract

Analysis of dyslexia is an interesting research area in medical field through MRI scans and eye ball tracking etc and also it is very interesting research in machine learning through different models. Dyslexia is a learning difficulty that affects the ability of reading, writing, listening and learning in children and adults. It is mostly misclassified as disorder but it is basically a difficulty. Many research’s in detecting dyslexic or non-dyslexic patients in many different ways were carried out by researchers, however the focus of our current study is to classify the dyslexia and further do the subgroup analysis of dyslexic patients using machine learning algorithms. Machine learning is a technique used in computer science in which models automatically learns the way humans learn. They learn from the data given to them. Machine learning algorithms are used mostly for prediction. In the current scope of research, for this purpose Different assessment techniques are used to assess students based on skills like reading, writing and sounds. The score of these assessments are further used for classification and analysis of subgroups in dyslexic patients. Supervised and un-supervised both techniques are used for classification and subgroup analysis. For classification purpose both binary class and multiclass classification is done. For binary classification purpose neural network, for multiclass classification svm and for clustering subgroup analysis k-means algorithm is used. In binary classification dyslexic and non- dyslexic patients have been classified. In multiclass classification different classes are classified and after this clustering is done using k-means algorithm to make different groups of different dyslexic level.