A New Classification Model Fuzzy-Genetic Algorithm for Detection of learning disability of Dyslexia in Secondary School Students

Background and Objective Learning disability is a neurological disorder. Simply, learning disabilities result from a person's misunderstanding of the brain. Children with learning disabilities are more accurate and intelligent than their peers, but they may have difficulty in reading, writing, pronouncing, concentrating, reasoning, recalling, or organizing information. Reading is the most basic and essential tool of education. Because by acquiring this skill, one will be able to acquire the necessary information in the affairs of life. The advancement of science in the present century is so rapid that reading is one of the most important tools for understanding today's world. One can learn the results of others' research and studies in a short period of time. Reading is a complex process that involves many different components. Learning disability is very common in childhood. The most important disability is reading disorder which is related to reading skills. Among the skills a student learns in school, reading is especially important. Meanwhile, there are students in higher grades whose reading progress is significantly lower than the standard level compared to their calendar age. This research represents a hybrid scoring model using genetic algorithm and fuzzy set theory to manage uncertainty in diagnosis of reading disability. Methods: For this, fuzzy classification models were applied for diagnosis of the reading disability. In the fuzzy system, the knowledge was extracted from a group of experts who were teachers and specialists. In the proposed model, the knowledge of experts was automatically extracted using the learning process of the Genetic algorithm. A dataset of 260 girl students was collected from the Marefat High school in the Alborz province in the years of 1394 and 1395. The performance of the proposed model was investigated using the ROC curve analysis. Findings: The results show efficiency of the fuzzy classification model was increased to 98.51% after the rule learning with the Genetic algorithm. The proposed fuzzy classifier models uncertainty in the knowledge of expert to improve students’ progress. Conclusion: The results of this algorithm show that compared to several other methods, the fuzzy-genetic combination method performs better than other methods. The results of the performance characteristic curve also prove this. Comparing the efficiency of the system and its analysis using ROC shows that fuzzy classification system is abl