Minería de Datos Educativa aplicada al descubrimiento de patrones en registros académicos de estudiantes de Ingeniería con asignaturas relacionadas a las matemáticas.

Worldwide, academic desertion is a phenomenon that persists in all higher education institutions. In Ecuador, according to a report presented by UNESCO, there is a 40% university student dropout rate; in some institutions it has been shown that there are various causes of dropout, such as social, academic, demographic, economic, pedagogical factors, among others. Although it has been tried, through different techniques, to reduce the dropout rates, it has been identified that these rates have increased due to the University admission program implemented by the Government. In addition, there is low interest in students with subjects related to Mathematics in Engineering careers since, according to the results obtained by the "Ser Bachiller" test, only an "Elementary" level is reached in this area, adding the rural sector as the most affected. By virtue of the above, the main objective of this Degree Project is to apply Educational Data Mining techniques to identify patterns in the academic records of students with subjects related to mathematics, in the Systems/Computer Engineering Career of the National University of Loja, which was carried out based on the phases of the CRISP-DM Methodology, which were developed through Excel, Google Colab and through the use of Python libraries such as Scikit-Learn and Pandas. The academic records were provided by the Directorate of Telecommunications and Information UTI, which were prepared for the topic of study through Python. During this process, the K-nearest neighbor technique was used to fill in some data, obtaining as output a model with which the different algorithms were applied, one for each cycle, to obtain the results of this TT. For the discovery of patterns, descriptive techniques were used, which were executed through: Association Rules (A-priori) and Clustering (K-means), implemented in three tools, Weka, RapidMiner and Python. Finally, through its execution and with the information obtained by each environment, factors such as not working, urban sector, male gender, single, are identified as the most predominant among the grouping of social, demographic and academic variables and from them, an action plan is proposed at a general level, with strategies that support the corresponding authorities in decision making, in order to reduce student desertion rates in the Systems/Computer Engineering Career. Keywords: Academic Record, Mathematics Subject, Engineering, Educational Data Mining, Crisp-DM, Descriptive Techniques, Weka, RapidMiner, Python.

• Autore principale: Villamagua Poma, Jhulissa Isabel (author)
• Natura: bachelorThesis
• Lingua: spa
• Pubblicazione: 2021
• Soggetti: <REGISTRO ACADÉMICO><ASIGNATURA DE MATEMÁTICAS><MINERÍA DE DATOS EDUCATIVA>< CRISP-DM>< TÉCNICAS DESCRIPTIVAS>< WEKA>< RAPIDMINER>< PYTHON>
• Accesso online: https://dspace.unl.edu.ec/jspui/handle/123456789/24426