A CLASSIFICATION MODEL FOR FACTORS AFFECTING ACADEMIC PERFORMANCES OF HEARING-IMPAIRED STUDENTS USING MACHINE LEARNING (IN CASE OF ADDIS ABABA EDUCATION OFFICE)

Abstract

Hearing is important for speaking and verbal development, communicate, and learning. Children with hearing loss or challenges with auditory processing continue to be an underserved and misunderstood society. The federal democratic republic of Ethiopia's education and training policy encourages Special Education for children with Special Needs in general, and hearing-impaired students in particular. Hearing-impaired children may face a variety of issues during the learning process as a result of their impairment, including isolation, low self-esteem, and learning difficulties. As a result, proper technological aid is required to improve the learning environment for hearing-impaired students. The outcomes of the study will aid educational planners, teachers, parents, and others in better comprehending the issue of hearing impaired education. The purpose of this research is to apply machine learning to classify factors that affect the academic performance of hearing-impaired pupils. Utilize a number of machine learning algorithms. Machine learning makes it possible to quickly create a classification model. We use the percentage split 70/30,60/40,90/10 ratio and we also use the data split 80/20 ratio to build classification models for each classifier algorithm. As an outcome of the experiment, we chose 80/20 split because it provided us with better accuracy. The factors affecting the academic performance of hearing-impaired students are using machine learning classification algorithms such as Support vector machine, Random forest, Decision Tree, GB, AdaBoost, XGB, and KNN. To assess an algorithm's performance, the confusion matrix, accuracy, precision, recall, and f1 have all been utilized. By 99.44 percent, the Gradient Boosting classification technique outperforms the other algorithms.