School of Computer Science and Engineering (Apr – May 2022)
Journal Paper:
1. Dutta, J. and Roy, S. (2022),OccupancySense:Context-basedIndoor Occupancy Detection &Prediction Using CatBoostModel,,Applied Soft Computing, Elsevier, Vol. 119, p. 108536,ISSN 1568-4946, DOI: https://doi.org/10.1016/j.asoc.2022.108536. (IF: 6.725).
Abstract:
Occupancy detection and prediction are well-established research problems where desired higher accuracy can be achieved with our proposed non-intrusive OccupancySense model which detects human presence and predicts indoor occupancy count by the fusion of IoT based indoor air quality data along with static and dynamic context data. This data fusion helps us to achieve higher forecasting accuracy along with the integration of state of the art gradient boosting based categorical features supported by the CatBoost algorithm. The speciality of this model is that it is non-intrusive and has very high predictive power outperforming other state-of-the-art models.
2. Priyadarsini, M., Bera, P., Das, S.K. and Rahman, A. (2022), A Security Enforcement Framework for SDN Controller Using Game Theoretic Approach, IEEE Transactions on Dependable ad Secure Computing (TDSC), DOI: 10.1109/TDSC.2022.3158690 (IF:7.329).
Abstract:
This paper presents an effective security enforcement framework for proactive prevention of potential attacks on SDN controller. First, based on a signalling game approach, we design a trust-based controller attack detection (TCAD) model that calculates the trust value of each incoming packet to take necessary action. Next, we propose a risk-based attack prevention (RAP) model that detects and filters malicious traffic flows in the network. Finally, we evaluate our proposed security enforcement framework on different scenarios with varying traffic requirements and by injecting attacks based on STRIDE model. Experimental results show 95% accuracy in the potential attack detection and prevention.
Conference Paper:
1. Maity A., Roy S.G., Bhattacharjee S., Ghosh R., Choudhury A. R. and Pradhan C. (2022), Image Encryption using RABIN and Elliptic Curve Crypto Systems, In: Springer International Conference on Emerging Technology in Computer Engineering (ICETCE), Jaipur, India, pages 691-703, 2022.
Abstract:
This paper has used two layers of each encryption and decryption thus providing two layers of security. Here we have worked with different images so firstly, the image has been encrypted using Rabin Cryptography followed by ECC Cryptography. Similarly while decrypting, the decryption was first done using ECC and then Rabin Cryptography. We have tested this approach using different datasets and have received great accuracy. With this approach it can provide images with a double layer of security efficiently while sharing images.
Book Chapter:
1. Bhattacharjee, P. and Mitra, P. (2022). Density-Based Mining Algorithms for Dynamic Data: An Incremental Approach. In: Dash, S.R., Lenka, M.R., Li, KC., Villatoro-Tello, E. (eds) Intelligent Technologies: Concepts, Applications, and Future Directions. Studies in Computational Intelligence, vol. 1028, pp. 313-335, Springer, Singapore. https://doi.org/10.1007/978-981-19-1021-0_13
Abstract:
Traditionally, data mining algorithms deal with static data. The major bottlenecks with this class of algorithms include redundant computation and higher latency along with increased consumption of available resources. Given the essence of dealing with dynamic data, this work focuses on designing incremental mining algorithms, specifically in the field of density-based clustering and outlier detection. The primary reason being density-based algorithms shows robustness in finding clusters of varying granularity or extracting outliers from variable density regions. We propose incremental extensions to two density-based clustering algorithms: MBSCAN (iMass), SNN-DBSCAN (BISDBadd, BISDBdel) and an outlier detection algorithm KNNOD (KAGO).
2. Das C., Sahoo A. K. and Pradhan C. (2022), Multicriteria-Based Entertainment Recommender System using Clustering Approach in Advanced Analytics and Deep Learning Models, A. Mire, S. Malik and A. K. Tyagi (Eds.), Wiley, pp. 33 – 63.
Abstract:
The first recommender system was based on single criteria, known as single-criteria recommender system. But in real-world scenario for recommendation, a model needs to look at more than one criterion. So, the multi-criteria recommender concept came in the picture. Here, we have seen some innovative ideas, approaches, and methods, which are applied to a multi-criteria recommender system more efficiently and effectively. We have also seen how these new and innovative approaches give better result compared to the conventional recommender systems. Here, we have talked about many different approaches of multi-criteria recommendation techniques done by various researchers around the globe.