Web Application Based on Neural Networks for the Detection of Students at Risk of Academic Desertion
Web Application Based on Neural Networks for the Detection of Students at Risk of Academic Desertion

Summary/Abstract: The objective of the study is to improve the prediction of academic desertion by means of a web application based on neural networks in a private university in Trujillo during 2023 The study employed and applied pure experimental design. The population comprised all the faculty's academic processes, with a sample size of 60 processes. The data collection technique was direct observation, and the instrument was an observation sheet. The web application was developed using Python 3 and the Flask framework, with MySQL as the database manager and Extreme Programming (XP) as the software development methodology. For the descriptive analysis, Microsoft Excel 2019 was used, while Jamovi 2.3.28 was utilized for inferential analysis. The results showed a reduction of 123 seconds in the average time for academic data collection, a reduction of 848 seconds in the average prediction time for identifying students at risk of academic dropout, and a 3% increase in the identification rate of at-risk students. In conclusion, the use of a web application based on neural networks significantly improves the prediction of academic dropout.

• Details
• Contents