Transportation, Civil and Structural Engineering, Waste Management and Disposal, Urban Studies
18
Scopus Publications
Scopus Publications
Harnessing machine learning models and explainable AI to understand MOOC continuance intention Vinod Sharma, Yogesh Mahajan, Manohar Kapse, Saikat Deb Information Discovery and Delivery, 2026 Purpose This study aims to investigate factors that influence individuals’ continuance intention to use massive open online courses (MOOCs) by using machine learning models. Design/methodology/approach Data was collected from 702 MOOC users from major metropolitan cities in India through a network-based sampling and recruitment via various social media outlets (e.g. LinkedIn and Facebook). Various machine learning algorithms along with explainable artificial intelligence (XAI) were employed using Python PyCaret. Findings Results confirm that pedagogy value, content value, interface value, ubiquity value, teacher presence and learning satisfaction have positive effects on the continuous intention of MOOCs. Furthermore, pedagogy value is a chief driving force of CI of MOOCs. XAI helps clarify intricate patterns in learner data, thus allowing more appropriate interventions. Practical implications The findings of the study would be useful for MOOC developers to formulate better value propositions for ensuring a sustainable business and higher growth rate in the market. Originality/value This study bridges the gap in the existing literature by providing a novel approach. To the best of the authors’ knowledge, this is the first study using machine learning and XAI earlier in identifying factors leading to continuous intentions of MOOCs, so this research adds value to the existing method of exploring factors enhancing retention rates among learners.
MOOC continuance: investigating the factors that keep learners engaged using PLS-SEM Vinod Sharma, Yogesh Mahajan, Sonali Bhattacharya, Saikat Deb, Manohar Kapse International Journal of Information and Learning Technology, 2026 Purpose Although Massive Open Online Courses (MOOCs) have experienced unprecedented growth globally, their popularity presents two of education’s greatest challenges today: low completion rates and user retention. This study aims to examine the various factors that influence the continuous usage of MOOCs amongst learners by integrating two theoretical models: The Expectation-Confirmation Model in Information Systems (ECM-IS) and Task-Technology Fit (TTF). The study assesses the impact of information quality, system quality, service quality, confirmation, perceived usefulness and learning satisfaction on long-term continued engagement in MOOCs. Design/methodology/approach Data were obtained from 377 respondents through a structured survey and analysed using partial least squares structural equation modelling (PLS-SEM) to identify factors determining success of MOOC. Findings Findings suggest that information and system quality strongly affect confirmation with subsequent effects on perceived usefulness and learning satisfaction, key antecedents of continuous usage intention. It also indicates that the matching of MOOC platform features with users’ learning tasks (TTF) positively influences satisfaction in learning by promoting subsequent long-term engagement. Originality/value This study provides some theoretical implications on the understanding of MOOC retention and some practical recommendations for MOOC providers in terms of content, system stability as well as support questionnaire adaptability to enhance user experience. Overall, the derived effectuation will offer a deeper understanding of the interconnection of variables that influence persistent usage with implications for practical strategies for sustaining learners in online education platforms.
Internal Carbon Pricing: Paving the Way for Low Carbon Transition in Indian Businesses Water and Energy International, 2025
Reuse of Brick Waste in the Construction Industry Saikat Deb, Mriganka Mazumdar, Rakesh A. Afre Journal of Mines Metals and Fuels, 2024 Demolition work produces a lot of waste or demolition waste made up of different materials like concrete, wood, metals, bricks, glass, plastics and asphalt. To maintain environmental sustainability this waste must be managed. This study offers a sustainable solution for the construction industry by providing a thorough analysis of demolition waste management with a focus on the reuse of brick aggregate and brick dust. Samples from a controlled demolition site were collected for the study. The brick samples from the demolition site were found through laboratory testing to have greater compressive strength than regular bricks, making them an acceptable building material. Brick dust was also discovered to be a superior void filler. These waste materials were used to create lean concrete, which was stronger than regular concrete and thus suitable for building. Cost comparisons revealed significant cost savings, making this strategy appealing from an economic standpoint. The study emphasises how using bricks made from demolition debris could lower carbon emissions. Responsible demolition waste management can also prevent priceless land resources from being turned into landfills, enhance soil quality and advance in building a circular economy. In conclusion, improving demolition waste management, particularly by recycling brick waste, offers a long-term solution for the building sector. It contributes to a greener and more sustainable future by minimising environmental impact, conserving resources, lowering construction costs and promoting a circular economy.
The Role of GPT and Data Fusion in improving Disaster Prediction Saikat Deb, Mriganka Mazumdar Disaster Advances, 2024 The application of data fusion for disaster prediction using GPT (Generative Pre-trained Transformer) is covered in this study. Data fusion is the process of merging information from several sources including social media, sensor data, and simulation data, in order to increase the precision of catastrophe prediction models. Through the utilization of GPT, an artificial intelligence language model, data fusion enables a thorough examination and amalgamation of data from disparate fields, hence producing more resilient forecasts. GPT models may be used to recognize geographical descriptions from social media postings and identify cell kinds using information about marker genes. Proactive communication with impacted people in times of catastrophe is made possible by the integration of GPT with social media monitoring. GPT models may significantly enhance disaster preparedness, response, prediction, and recovery by gathering pertinent data from many sources and modelling various situations.
Assessment of Changes in Brahmaputra River Course at the Pagladia Confluence Point using Remote Sensing and GIS Techniques Mriganka Mazumdar, Saikat Deb Disaster Advances, 2023 The river Brahmaputra is a large alluvial river that is prone to frequent bank erosion and channel pattern changes, leading to significant shifts in its course. This study aimed to analyze these changes along a 56-kilometer stretch of the river using a combined approach of remote sensing and GIS techniques. This study utilized USGS and Landsat 8 satellite imagery to map the river's channel configuration from 1985 to 2022, providing valuable insights into the river's morphology and the stability of its banks. Additionally, the analysis provided information on changes in the river's main channel which can help in predicting future behavior and mitigating the impact of these changes. The findings of this study have significant implications for river management, allowing for informed decision-making and improved strategies for protecting communities and infrastructure located along the river's course.
Variability Of the perception about safety, comfort, accessibility and reliability of city bus service across different users groups International Journal of Civil Engineering and Technology, 2018
Publications
Das, Jayanta Kumar, Saikat Deb, and Biswadeep Bharali. "Prediction of aggregate impact values and aggregate crushing values using light compaction Journal of Applied Engineering Sciences 11.2 (2021): 93-100.