Machine Learning Model for Applicability of Hybrid Learning in Practical Laboratory
Chaman Verma · Procedia Computer Science · 2024
Significant changes have been observed from 2019 in the COVID period as study styles shifted from traditional to hybrid learning. This paper predicted the applicability of the hybrid learning mode of education in the programming labs for the informatics faculty. It explored novel vital features that predicted the applicability of hybrid learning in the practical labs of the informatics domain. For this, the Classification and Regression Tree (CART) identified the relevance of hybrid education with an accuracy of 71.7%. The confusion matrix of the CART model has 67.3% correctness (sensitivity) for “Yes”, and 76% correctness (specificity) for “No” during validation. The false negative rate (type II error) is 32.7%, and the false positive rate (type I error) is 24%. The AUC value of the CART model is 0.68. Some obstacles were identified, such as isolation from the natural study environment, reduced quality of interaction with other faculty, and turning off the mic or camera during labs. Also, it was found that the university initiated and provided support to create less stress and ease of communication. Adopting the Hyrbid learning mode is unsuitable for handling group tasks or projects in the laboratory. Also, the hybrid learning mode proved a vital mode of study in practical labs as online learning with a safe environment.