Vishwakarma University, Pune | Impactful Research Projects

Name of the Faculty :	Noshir Tarapore
Designation :	Assistant Professor
E-mail :	noshir.tarapore@vupune.ac.in
Faculty :	Science and Technology
Department :	Computer Engineering
LinkedIn :	https://www.linkedin.com/in/noshirtarapore
Google Scholar URL :	https://scholar.google.com/citations?view_op=list_works&hl=en&authuser=1&user=tN8KFEcAAAAJ&gmla=AJsN
ResearchGate :	https://www.researchgate.net/profile/Noshir-Tarapore
ORCID ID :	https://orcid.org/0000-0002-7303-6042
Research Area :	Support vector machine, intrusion detection system, parallel computing, speedup

Research Summary :

His research interests include working on Machine Learning projects that have applicability in the Network Security domain. This includes Performance Improvement Strategies in Intrusion Detection Systems using Supervised Learning Techniques. Improving the performance in terms of speedup is one such aspect. Intrusion detection can be treated as a pattern recognition problem which distinguishes between network attacks and normal network behavior. Training a Support Vector Machine (SVM) to use it for classification, to predict whether an attack has taken place or not, still remains computationally very intensive. Much research has been done to accelerate training time, and thus there is a need to derive an efficient scalable SVM solver which can use the computational power of the Parallel computing paradigm to reduce the training time of an SVM. Neural networks, although slower, could also be used for intrusion detection. His inter disciplinary research interests include algorithms that could be used in Economics or other financial domains such as stock markets. Algorithms that could predict times to buy and sell particular stocks would be an area of interest. Models that could be trained using different financial parameters to aid humans in decision making would be of particular interest. Taking things a step further, would be to reduce the input features to such a model to reduce the time required to train the model. Only features which are critical to the model should be retained, relatively redundant ones need to be discarded.