Transparent Peer Review By Scholar9
ENHANCING STUDENT ENGAGEMENT TRACKING DURING ONLINE LEARNING USING DEEP LEARNING
Abstract
Changes in virtual learning environment due to Covid-19 epidemic have faced challenges in effective monitoring of student engagement during online classes. This study proposes a novel hybrid deep learning approach using a bagging dress of CNN1D and Resanet1D architecture to automatically detect student engagement. The model is trained on the DAISEE dataset, with several engagement labels, and addressing data imbalance using SMOTE technology. Similar deficiency is obtained through monotonal value decomposition (SVD) to increase model efficiency. The hybrid model standalone displays better accuracy than the CNN1D and Resanet1D models. The major app modules include detecting real -time engagement through webcam and video analysis. Experimental results indicate that the proposed hybrid enclosed method effectively identifies various engagement levels, offering a scalable solution to increase the quality of online education.
Niravkumar K Patel Reviewer
Revision Required
Hello Researcher,
I like the research paper, but I think it should be improved a lot. I can't see any depth process to handle the online and hybrid model with proper techniques.
I can see several issues in using the algorithms.
If we can use one of these algorithms, then it will be very good for this learning process because it is one of the best algorithms to handle the online process with accuracy and security. This will have a lot of impact on students in the future.
Data Encryption & Decryption Pseudocode (AES-like)
1. Key Generation
plaintext
CopyEdit
function generateKey():
key = random 256-bit value
return key
2. Encryption Process
plaintext
CopyEdit
function encrypt(data, key):
iv = generateRandomIV() // Initialization Vector
cipher = AES_Encrypt(data, key, iv)
encryptedData = iv + cipher // prepend IV for decryption
return encryptedData
3. Decryption Process
plaintext
CopyEdit
function decrypt(encryptedData, key):
iv = extractFirstBlock(encryptedData)
cipher = extractRemainingBlocks(encryptedData)
originalData = AES_Decrypt(cipher, key, iv)
return originalData
We can use this kind of process in the use of CNN1D algorithms to improve the process. The overall research is good. Thanks for giving me the chance to review.
Niravkumar K Patel Reviewer
23 Jun 2025 10:32 AM
Approved
Hello Researcher,
I hope you are doing well. I have given several comments to improve this research. It will be very helpful to correct.
IJ Publication Publisher
Dear sir,
We have conveyed the comments for revision to the author. Thank you for your valuable review comments.
Niravkumar K Patel Reviewer