Bibliographic citations
G., A., Valencia, J. (2025). Aplicación web de retroalimentación de sesiones de clase y su eficacia en el rendimiento académico [Tesis, Universidad Privada Antenor Orrego]. https://hdl.handle.net/20.500.12759/58112
G., A., Valencia, J. Aplicación web de retroalimentación de sesiones de clase y su eficacia en el rendimiento académico [Tesis]. PE: Universidad Privada Antenor Orrego; 2025. https://hdl.handle.net/20.500.12759/58112
@misc{renati/1709167,
title = "Aplicación web de retroalimentación de sesiones de clase y su eficacia en el rendimiento académico",
author = "Valencia Barriga, José Javier",
publisher = "Universidad Privada Antenor Orrego",
year = "2025"
}
This thesis report, entitled “Web application for class session feedback and its effectiveness on academic performance”, addresses the problem of insufficient feedback in the Computer and Systems Engineering program of study at the Universidad Privada Antenor Orrego, which negatively affects the academic performance of students by not consolidating the knowledge acquired. This research, conducted by Ana Cristina G. Otoya Sifuentes and José Javier Valencia Barriga, aims to evaluate the effectiveness of a web application that integrates Speech Service and GPT-3, facilitating feedback through the transcription of classes, automatic summaries and an interactive chat. The comparative experiment between an experimental group and a control group revealed a remarkable increase in the academic performance of the group that used the application, obtaining an improvement of 30.43% from pre-test to post-test, in contrast to the increase of only 5.26% of the control group. As for the goal of achieving the maximum score, the experimental group advanced 17.5% towards this objective, while the control group obtained only 3.1%. Likewise, the usability of the application was evaluated according to ISO/IEC 25000, obtaining favorable results in all its sub-characteristics, achieving an overall rating of “acceptable”. On the other hand, the transcription tests obtained a word error rate (WER) of only 3.55%. Finally, the generated summaries achieved excellent quality according to ROUGE metrics: ROUGE-1 with 69% accuracy, 64% recall and 66% F-measure; ROUGE-2 with 45% accuracy, 42% recall and 44% F-measure; and ROUGE-L with 65% accuracy, 61% recall and 63% F-measure. In conclusion, this application is presented as an effective tool to improve academic performance, offering an accessible experience and optimizing the feedback process through artificial intelligence
This item is licensed under a Creative Commons License