Un modelo basado en el Clasificador Naïve Bayes para la evaluación del desempeño docente
| Autor: | Juana Canul Reich, Lourdes Y. Margain Fuentes, Guadalupe Gutiérrez Esparza, Tania Aglaé Ramírez del Real |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Higher education
business.industry Computer science Process (engineering) 05 social sciences 050301 education Confusion matrix 050801 communication & media studies Machine learning computer.software_genre Test (assessment) Naive Bayes classifier 0508 media and communications Classifier (linguistics) ComputingMilieux_COMPUTERSANDEDUCATION General Earth and Planetary Sciences Artificial intelligence Set (psychology) business 0503 education computer Mobile device General Environmental Science |
| Zdroj: | RIED. Revista Iberoamericana de Educación a Distancia. 20:293 |
| ISSN: | 1390-3306 1138-2783 |
| DOI: | 10.5944/ried.20.2.17717 |
| Popis: | The evaluation of teacher performance is an important measurement process in Mexico's higher education institutions and around the world, because it allows feedback on the teacher’s performance to detect improvements in classes and propose strategies for the benefit of students' education. This paper describes the development and evaluation of a proposed computational model called SocialMining, which is based on the classifier algorithm Naive Bayes to support the analysis of students' opinions from the process of teachers' performance evaluation, which is carried out through mobile devices. The mobile device revolutionizes processes in education; the proposal considers the use of this technology for the collection of data, taking advantage of processing capacity and acceptance by students in the process of education and learning. It also describes the development of a set of relevant affective terms of the teacher evaluation called corpus of subjectivity, which supports the Naive Bayes algorithm to classify students' comments within the classes: positive, negative and neutral. To measure the comments classification performance of the SocialMining Computational Model, metrics such as the confusion matrix, precision, sensitivity, specificity and the ROC curve are used. Likewise, a study case is presented, which gathers new comments from students of the Polytechnic University of Aguascalientes (Mexico), in order to test the classification process performance of the proposed model. The results show that SocialMining Computational Model is feasible to be implemented in institutions to support Teacher Performance Assessment. Besides, our results show that Naive Bayes can obtain a classification percentage very similar to those reported in recent works with related algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|