Aplicação de técnicas de pré-processamento e agrupamento na base de dados do aplicativo michelzinho

Abstract

Autistic Spectrum Disorder (ASD) is characterized by severe impairment in communication, interaction and behavioral learning, presenting the patient, in addition to social difficulties, with repetitive and stereotyped behavior. These characteristics impact the social interaction of these individuals, including the detection of facial expressions and recognition of basic emotions, disconnecting the patient from the social environment. However, using serious games to help teach and improve skills, psychology has achieved good results in treatment. In this way, the MICHELZINHO application aims to detect and recognize emotions, culminating in the development of skills in a fun way, in addition to generating a robust database to transmit new knowledge about the pathology to ASD specialists. Thus, this work was developed with the objective of analyzing the applicationŠs database, using data mining through computational algorithms. For this, a tool was developed, in python, used in the pre-processing of the data, in order to group them through three different algorithms. The results were validated using the simpliĄed silhouette, indicating that the proposed tool contributed to the improvement of emotion detection and recognition skills, despite the high dimensionality of the data after pre-processing