A Effective Method for Predicting the Dyslexia by Applying Ensemble Technique

Dyslexia is a condition where a person will face difficulties in certain tasks including reading, writing, speaking, and identifying sounds. Around 10% of people globally struggle with this issue. The most important step in preventing dyslexia is early identification. There are several ways to estimate the risk of dyslexia, where we have developed a model which allows the user to specify their language vocabulary, memory, speed, visual discrimination, audio discrimination test results. The model will determine the user’s individual risk of dyslexia after receiving input from the user. The approach we used included data preparation, data preprocessing, model training, model testing, and model construction. Predicting Risk of Dyslexia-PLOS ONE dataset is used. Dyslexia can be identified using machine learning classification techniques like Decision Trees, Random Forests, and Support Vector Machines. When compared to individual classification strategies, the ensemble technique in the proposed work predicts the risk of dyslexia with a better degree of accuracy. Here, we consider integrating GridSearch CV, Support Vector Machine, and Random Forest. Accuracy, precision, recall, and F1-score were taken into consideration as outcome measures.