An Efficient Machine Learning Model for Prediction of Dyslexia from Eye Fixation Events

Abstract

Dyslexia is a type of learning disability in which a person has trouble spelling and reading words fluently. Dyslexia is not curable, but with the correct remedial help, dyslexics can achieve great success in school and in life. Eye movement patterns during the reading process can provide a deeper knowledge of dyslexia-related reading difficulties. Eye movements can be recorded with an eye-tracker, and the relationship between how eyes move in proportion to the words they read can be deduced. Based on statistical measurements, a collection of binocular fixation and saccade properties were derived from raw eye tracking data in this study. Based on statistical measurements, a collection of binocular fixation and saccade properties were derived from raw eye tracking data in this study. Machine learning algorithms such as the Random Forest Classifier (RF), the Support Vector Machine (SVM) for classification, and the K-Nearest Neighbor (KNN) for prediction of dyslexia were investigated to provide classification models for dyslexia prediction. In comparison to SVM and RF, KNN provided 95 percent accuracy over a small feature set associated to fixations and saccades. These characteristics of the eyes can be exploited to design screening tools for dyslexia prediction. Early detection of dyslexia can assist children in receiving treatment, allowing them to achieve academic success.