Adaptive sequential recommender system for disruptive personalized dyslexia intervention

Abstract

Computer-assisted programs have been proposed as a solution to facilitate the systematic application of scientific knowledge in individualised and progressively more intensive interventions for children with dyslexia. These programs do not always meet the unique needs of each child, highlighting the urgent need for improved adaptive technology-based solutions that can provide tailored support throughout the learning process. This article presents a new adaptive sequential guidance system for personalised dyslexia intervention. The system introduces a number of key innovations: a dynamic word generator that creates phonetically modified words and pseudowords from seed words, a three-dimensional matrix structure (E, W,and F) for effective word difficulty manage- ment and user performance, and an algorithm of recommendations based on stochastic gratising semi-monthly matrix factors. The system uses a heuristic initiation process to reduce cold start problems and uses an extension technique to detect difficulties in certain derived words. Furthermore, the text introduces the concept of ”virtual children” based on Bayesian Knowledge Tracking, which allows for comprehensive testing and optimisation of systems before real implementation. The proposed system offers a unique approach to dyslexia intervention by dynamically adapting word difficulties based on individual user performance, ensuring that each child remains in its optimal learning area. The main results conclude that the use of thermal maps and 3D visualization F-matrix allows each user to identify specific difficulty areas, promoting more targeted intervention; additionally, extensive testing shows that the system is sufficiently robust to reduce error rates in several trials; and the parametric study reveals the system’s ability to adapt using adjustable parameters such as success and failure modulators and increasing factors; furthermore, the system demonstrates a strong adaptability of individual users.