Regresar

Creation of an Intelligent System to Support the Therapy Process in Children with ADHD

Abstract:

This article is a proposal for an intelligent system that facilitates the therapy process for children with Attention Deficit Hyperactivity Disorder - ADHD. This project focuses on the analysis, design, and implementation of an intelligent architecture, which uses machine learning techniques for a therapeutic robot to help develop new tests and other resources to support the treatment of ADHD. Its general purpose is to identify common patterns in the behavior of children with ADHD, between 6 and 9 years old during the performance of homework to help the therapist diagnose and pbkp_redict the future behavior of children in this area. A system has been proposed whose development and implementation is carried out by a cloud computing platform, taking advantage of all its benefits such as low latency, unlimited storage, functionalities to deploy a project based on artificial intelligence and data security. Therefore, using these functionalities from the platform, a machine learning model has been deployed. This model is a binary classification since it groups the results according to the diagnosis (ADHD or not ADHD). However, what this model provides are the rules that compare the data obtained from the robot’s sensors to the set of results expected to be obtained (ADHD or not ADHD). To obtain these rules, an algorithm already developed by this platform and a set of data will be used. This first part is known as model training, in which the model has been built. Once the rules are obtained, another set of data will be used for testing. In this second part, the model will be able to identify whether the new data entered matches the criteria of a child with ADHD or with a typical development. The data processing is carried out from the cloud platform, offering data availability and accessibility at all times. The results of the correlation between the obtained data and the pbkp_redicted diagnosis showed remarkable results.