Exploration on Learning Disorder using Machine Learning

A particular type of learning difficulty called dyslexia is most prevalent in children under age of twelve. Being not able to compete with their friends might make them feel hopeless, unmotivated to complete tasks, or depressed. Having trouble keeping up with youngsters that are different when it comes to terminology, numbers, and symbols. This research has made an effort to examine a number of machine learning algorithms for early detection and group classification. The evaluation techniques employed in this project include ensemble, neural network, SVM, and Naive Bayes. The assessment and assessment were conducted using MATLAB-2021. The EEG dataset encompasses 65 young people's facts, organized into 4 classes and 12 traits, of which 22 have been tested and 43 have been trained. In a 10-fold validation with 67% data split, the confusion matrix demonstrated that the SVM was the most effective version for both training and testing, with an accuracy of 52.5% for tests, and an excellent validation accuracy of 58.45% for ensembles. SVM has a 77% specificity, however ensemble has a 71% specificity. This study demonstrates that the dataset's performance indicates that the categorization parameter is dependent on gender, age, and mixed handedness of the group.