Desarrollo de una aplicación móvil para clasificación de imágenes de Alzheimer mediante aprendizaje automático

Abstract

According to the policies established by the agencies in charge of guaranteeing the quality of Higher Education in Ecuador, there is a need to implement mechanisms to eliminate barriers to access, permanence and completion of higher education studies; components related to the analysis of students' academic performance. In this sense, a Decision Support System (DSS) was developed contemplating the activities described in the Hephaestus and Kimball methodologies; which uses as a set of data information from the student welfare department (student data), the Human Resources department (faculty data) and the academic system of the institution (student grades). The proposed DSS generates consolidated information through statistical graphs and, using Pearson's correlational coefficient, calculates a measure of linear dependence between two of the variables studied (grade point average, marital status, gender, household income, source of study resources, type of household and type of housing) in order to determine the type and degree of association of the factors that influence student academic performance. Subsequently, using usability evaluation criteria of business intelligence systems, a survey was applied as a data collection instrument to the personnel involved in decision making; with the results, the Chi-square statistic was determined with a significance level of 5%, thus concluding that the implementation of a DSS optimizes time and human resources in the generation of consolidated reports that support the analysis of the indicators that influence academic performance. In future work, we recommend the inclusion of data sets that could be generated by the institution's subsystems such as research projects, outreach projects and publications, which will significantly enrich the DSS and thus support the decision-making process.