<![CDATA[The low initial reading ability of students in Bali is often due to uniform teaching and a lack of early detection of individual learning disabilities. This study aims to develop and test the effectiveness of a Deep Learning-based SIAGA (Early Reading Disorder Identification System) application to support instructional differentiation strategies. By using the Rapid Application Development (RAD) method. The research population is elementary and secondary school students in Buleleng Regency, Bali, and the research sample was taken by a purposive sampling technique. The data analysis technique is to measure the average results of UEQ, benchmarks, and comparative test results. The results of the test on 100 teachers showed that SIAGA had an identification accuracy rate of 96%, far exceeding the manual method, which only reached 42%. In terms of usability, this application obtained a System Usability Scale (SUS) score of 82.35 (categorized as Excellent) and an evaluation of the User Experience Questionnaire (UEQ), which shows excellence in terms of efficiency and novelty. This study concluded that SIAGA was effective in reducing identification time from 47.2 minutes to 14.6 minutes per 10 students, while providing an accurate data foundation for teachers to implement differentiated and targeted learning interventions to detect students' early reading ability.]]>
Copyrights © 2026
