Navigation of a Research Study on Student Depression Incorporating Various Machine Learning Approaches

This paper tries to navigate the depression impact on the student’s academic performance using machine learning algorithms. Total five machine learning models have been processed to see the performance of the students with different depression level. From the analysis it is clear that students face a critical problem in their academic life due to this depression stress. The dataset for this study has been collected from Kaggle website. This dataset comprises 27898 data samples where 16335 students are facing with critical depression stage. The age of these students ranges from 18 to 59 years. And, the CGPA scale is taken as 0 to 10 and the academic pressure is taken as 0 to 5. P value has been justified to remove insignificant features for optimizing the accuracy of the machine learning algorithms. Before feature retraction the accuracy for Random Forest has been recorded as 84% whereas it is 81% for the Naive Bayes model. After feature selection these models showed a better performance which is recorded as 88% for the Naïve Bayes model and 100% for the Random Forest. On top of this, after standard scaling the parameters, the accuracy for five models have been raised to the utmost level 100%.