Classification and Recommendation Techniques for Effective Learning in Flipped Classroom Pedagogy

Shaw, Rabi (2023) Classification and Recommendation Techniques for Effective Learning in Flipped Classroom Pedagogy. PhD thesis.

[img]PDF (Restricted upto 17/04/2026)
Restricted to Repository staff only



Flipped learning (FL) is found to be an effective teaching methodology which is accomplished in two stages. In the first stage, students take instructions and learn from pre¬loaded lecture videos (out¬of¬class learning). In the second stage, students carry out various activities such as group discussion, think¬pair¬share, group quiz, etc. in presence of the instructor (in¬class learning). Therefore, students get enough time to brainstorm on the topic learnt from the pre¬loaded lecture. This new learning pedagogy offers quality learning for many students. However, this teaching methodology does not have provision to monitor students while taking lesson from pre-loaded lecture video unlike traditional classroom teaching. This may lead to severe learning incompetence for weak students. On the other side, we know brain’s activity is very vital in deciding cognitive ability. Psycho-cognitive state of weak students can help in understanding their difficulties towards a specific topic and it can play an important role in helping them out. To assess the cognitive state of the people, usage of EEG signal has drawn increased attention in the last few years, particularly in Brain¬Computer Interface design. Capturing EEG signal is a simple, non¬invasive, and commonly used technique for analyzing the functioning of the human brain. Therefore, brain activity during taking lesson can be captured by collecting EEG signals. Researchers have started using multi¬channel EEG headset to capture more fine details. However, students may not be fully attentive with wearing multi¬channel EEG headsets due to heavy weight. Capturing attention for multiple sources involves costly deployment of various equipments. Single channel headset is cost¬effective and easy to wear. This thesis mainly focuses on enhancing the learning ability of the learner by identifying the weak student and recommending the non¬attentive lecture videos in FL pedagogy. Brain signal or electroencephalogram (EEG) of students can be utilized to address these (monitoring in FL and wearing multi¬channel EEG headset) issues by exploiting classification and recommendation techniques using single channel EEG headset in this thesis. In this thesis, an experimental set¬up is developed to monitor the student in flipped learning by capturing the brain wave of individual students passively while they are engaged with lecture video. The siamese neural network is exploited to analyze captured brain waves (EEG signal) in order to classify the students into three categories (weak, good, outstanding) based on their attention level. The experimental result shows that the proposed siamese neural network based model outperforms other classification models. Feature extraction technique can be used to improve the accuracy of the classification techniques. Two feature extraction techniques are proposed in this thesis. A popular feature extraction technique called Local Binary Pattern (LBP) is adapted to extract unique features from collected electroencephalogram (EEG) signals. The unsupervised deep neural network technique called variational autoencoder (VAE) is also exploited to extract useful features from captured EEG signals. Finally, standard classification techniques are used to classify the attention level of students. Cognitive state of an individual student is analyzed using brain waves signals while taking instructions in the absence of an instructor. A recommendation technique termed as Lecture Video Recommendation in Flipped Learning (LRFL) is proposed for effective learning in flipped classroom. In this technique, the brain waves (Electroencephalogram (EEG)) signal is analyzed using unsupervised learning (clusters) techniques to group similar behaviors exhibited by student over video duration. Based on this analysis, proposed recommendation technique detects non¬attentive video and suggests for retaking the lesson. Results demonstrate the effectiveness of our recommender technique. Data acquired at our laboratory for research on flipped learning has been utilized in all experiments.

Item Type:Thesis (PhD)
Uncontrolled Keywords:Classification; Cognitive State (CS); Discrete Wavelet Transform (DWT); Electroencephalography (EEG); Flipped Learning (FL); Mel Frequency Cepstral Coefficient (MFCC); Recommendation System (RS); Siamese Neural Network (SNN); Variational autoencoder (VAE)
Subjects:Engineering and Technology > Computer and Information Science
Divisions: Engineering and Technology > Department of Computer Science Engineering
ID Code:10501
Deposited By:IR Staff BPCL
Deposited On:16 Apr 2024 12:45
Last Modified:16 Apr 2024 12:45
Supervisor(s):Patra, Bidyut Kumar

Repository Staff Only: item control page