Identification of student productivity challenges in Bangladesh:

Abstract

Natural Language Processing (NLP) approach is adopted in this research to identify the causes of students’ productivity problem in Bangladesh. To the best of the authors’ knowledge, the current study adopted a cross-sectional design where 3, 079 entries and two-columns questionnaires incorporating the five target attributes; ‘Drifting’, ‘Technoference’, ‘Laziness’, ‘Rushed’, and ‘Procrastinating’ were employed to reveal the underlying drivers of productivity interruption. Particularly, the study is based on surveys, interviews of learners, and the data gathered from social networks to consider the main difficulties that high school and university students meet at their studying process. In the variety of Deep Learning and Machine Learning techniques, namely Decision Trees (DT), Support Vector Machines (SVM), Logistic Regression (LR), Random Forest and Convolutional Neural Networks (CNN) algorithms are used for the dataset. The efficiency of these models is measured based on accuracy and here the accuracy of the CNN model is 99.58%. They perform well in predicting and classifying the productivity interruptions, hence depicting a use of the algorithms to address productivity issues among students. Therefore, these findings of the study are pertinent to educational policymaking, administration and practice in Bangladesh. Thus, achieving higher learning results, the primary causes negatively influencing student productivity are revealed to create specific measures and actions. Additionally, the study adds to the existing body of knowledge on the problems of education in developing nations and shows how NLP and machine learning can be applied for solving another tough issue concerning education.