Research 

Research Interest 

• Data Science

• Machine Learning

• Natural Language Processing

• Deep learning

Undergrad Research

My journey into "ConcentrateNet: A Deep Learning Architecture for Analyzing students’ Concentration in online courses through Webcam" began with a deep curiosity about Data Science and Machine Learning. My interest in data science developed from its straightforward, useful use in understanding and assessing real-world data, which enables me to derive conclusions with concrete implications for a range of circumstances and problems.  Making a robust and unbiased dataset from scratch was my first significant challenge, which helped me improve my data visualization and cleaning abilities. My technical knowledge was enhanced by creating three models that were able to forecast concentration levels with greater accuracy and by experimenting with deep learning and neural network approaches using image data. Moving forward, I aspire to apply these findings to develop more advanced tools that can enhance the effectiveness of online education, ultimately contributing to better learning outcomes and more personalized educational experiences.

ConcentrateNet: A Deep Learning Architecture for Analyzing Students’ Concentration in Online Courses Through Webcam

Abstract:

 Online learning is growing in popularity these days. As a result, students typically contribute millions of course-related responses to discussion forums and exchange some learning experiences. This study focuses on online courses offered through MOOC platforms and identifies the variables that affect students’ ability to stay focused. We suggest a unique method to address this issue by evaluating students’ levels of concentration using the CNN architecture, MobileNetV2, VGG16, ResNet50, and InceptionV3 models. Our goal is to determine whether the issue is with students’ concentration, the course material, or both. Measurement of concentration levels, evaluation of video data, comparison of model performances, and provision of class-based concentration levels (attentive, inattentive, and sleepy) are the goals of our research. The dataset underwent pre-processing, which included resizing for analysis, frame extraction, and annotation for classification. Our research offers educators insightful information that will help them to increase the overall efficacy of online learning. Furthermore, the study advances the area by offering a methodical technique for assessing and evaluating students’ concentration on online courses.

 Keywords: Concentration levels; CNN; MobileNetV2; VGG16; ResNet50; and InceptionV3