Using Machine Learning Predicting Satisfaction of Students towards Hybrid Learning

In this research, we provide a machine-learning approach with novel features for automatically predicting student satisfaction with hybrid learning. One hundred students in the informatics department served as the study’s primary samples. We modeled three classic machine-learning algorithms—Random Forest, Xtreme Gradient Boost (XGBoost), and Logistic Regression—to address the binary classification challenge of foreseeing students’ feelings of contentment about hybrid pedagogical approaches. The average score for Cronbach’s alpha test to assess the reliability of 31 separate features was 0.825. The XGBoost algorithm had the highest mean accuracy (97%) in predicting student satisfaction. Still, all three algorithms provided us with significant accuracy (One-way ANOVA also proved the identical significant accuracy of all three algorithms (p>0.5).The paper’s findings recommended that hybrid learning will be a long-term solution. Students are willing to adopt a hybrid learning mode in their studies as Online Theory and offline lab on alternate weekdays. Also, the results show that students were happy and satisfied with time management to attend two classes simultaneously. Also, they believed hybrid learning is very convenient for sickness and bad health. We also found that parents of students also liked hybrid learning mode in future severe pandemics like Covid-19, and students themselves committed that hybrid learning is the best solution for them as an online learning opportunity with safety.