Associate Professor Ph.D. (IIT Kanpur)
+91 (0)361 258 2620 arabin ◊ iitg ⋅ ac ⋅ in
Research Area: Distributions, models and Applications; Machine learning and Time Series Analysis ; Advanced Statistical Algorithms; System Biology, Biomedical Applications, Ensemble Learning.
Room No: E1-310
Personal Website | IRINS Profile
System Biology, Biomedical Applications, Ensemble Learning : Conditional Survival Predictions, Variable Section, Outlier detections are few problems relevant for the doctors and clinicians. Longitudinal, censored, competing risk are some of data patterns that are available not only in biology but also in different areas of science and technology. Any patient specific analysis or drug discovery often includes both phenotype and Genotype attributes in their clinical analysis. The research methodology in this direction not only includes accurate prediction of specific scenario but also equally emphases on biological interpretation of the proposed models.
Machine learning and Time Series Analysis : Accuracy in forecast for long term signal is a never ending problem. We focus on uncertainty estimation for forecast in multi-dimensional time series model through machine learning algorithm. Conformal interval prediction in multidimensional signal is a challenging problem.
Advanced Statistical Algorithms : I love to work building new algorithms that solves practical problems. I contributed in developing EM Algorithm, Advanced MCMC and its convergence (Gibbs, Slice, Simulated Annealing, HMC, NUTS etc). My interest includes applications on HMM, Genetic algorithms and all Deep learning algorithms.
Distributions, models and Applications : One of the most difficult problem in Statistics is choice of a distributions given a data. The problem becomes difficult when data is censored, heavy-tailed, discrete, high dimension or any combination of them.
Research Project Associate in Speech Signal, a joint project of BSNL, Department of Mathematics and Statistics, Department of Computer Science & Engineering and Department of Electrical Engineering.
Director at Medsolu OPC Pvt LTD from 2020.
1) MA 373/471 : Statistical Analysis of Financial Data, –
2) MA 226 : Monte Carlo Simulation –
3) MA 225 : Probability and Random Process, –
4) MA 598 : Mathematics of Financial Derivatives, –
5) MA 225 : Financial Engineering I –
6) MA 477 : Financial Risk Management, –
7) MA 691 : Statistical Data Analysis and Simaulation, –
8) MA 691 : Advanced Statistical Algorithm, –
9) MA 593 : Statistical Methods and Inference, –
11. Biplab Paul and Arabin Kumar Dey, "An EM algorithm for absolutely continuous Marshall-Olkin bivariate Pareto distribution with location and scale", Communications in Statistics - Simulation and Computation [2023] , https://doi.org/10.1080/03610918.2023.2293644
10. Shankhajyoti De, Arabin Kumar Dey and Deepak Kumar Gouda, "
Construction of Confidence Interval for a Univariate Stock Price Signal Predicted Through Long Short Term Memory Network
", Annals of Data Science , vol.9 (2) [2022]9. Sangita Dey, Arabin K Dey and Rajesh Kumar Mall, "
Modeling Long-term Groundwater Levels By Exploring Deep Bidirectional Long Short-Term Memory using Hydro-climatic Data", Water resources management ,pp.3395-3410, [2021]
8. Arabin Kumar Dey and Biplab Paul , "
7. Biplab Paul, Arabin Kumar Dey and Debasis Kundu, "Bayesian analysis of three parameter absolute continuous Marshall–Olkin bivariate Pareto distribution", Communications in Statistics: Case Studies, Data Analysis and Applications ,pp.57-68, [2017]
6. Arabin Kumar Dey, Sreangsu Acharyya and Raghav Somani, "
5. Kundu D. and Dey AK, "Discriminating between bivariate generalized ex- ponential and bivariate Weibull distributions", Chilean Journal of Statistics [2012]
4. Arabin K. Dey and D. Kundu, "Discriminating between the log-normal and log-logistic distributions", Communications in Statistics - Theory and Methods , vol.39 ,pp.280-292, [2010]
3. Arabin Kumar Dey and D. Kundu, "Discriminating between the Weibull and Log-normal distributions for type-II censored data", Statistics (to appear) [2010]
2. Arabin Kumar Dey and D. Kundu, "Discriminating among the log-normal, Weibull and generalized exponential distributions", IEEE Transactions on Reliability , vol.58 (3) ,pp.416-424, [2009]
1. D. Kundu and Arabin Kumar Dey, "Estimating the Parameters of the Marshall Olkin Bivariate Weibull Distribution by EM Algorithm", Computational Statistics and Data Analysis , vol.53 (4) ,pp.956-965, [2009]
M.Sc.
| 1 | Ritika Verma |
| 2 | Aastha Barnwal |
| 3 | Swamya |
Research Scholars
| 1 | Biplab Paul |
No Project Available!
|
1. Business Development new
Thanks Lakshit Sethia for completion of three applications containing three Businesses with valuation in Trillions. 2025-05-15 |
|
2. A Proposition To Remove Biometric Attendance And New Line Of Innovations new
Recently we completed a mobile application on obtaining attendance from CCTV videos. Research Innovation : The research proposes a image and video integration on Generative Adversarial Boosting Model to restore the obscure image and videos in attendance generation. Interested Students can join more on the research and application. CCTV Companies are welcome to contact for integration in real life integrations. 2025-05-15 |
|
3. Looking for collaboration with More on sequential Cobra new
First Sequential Cobra structure is available at https://arxiv.org/abs/2405.04539 with innovation in hyper-parameter optization to implemention of boosting to improve the accuracy in multi dimensional signal forecasting. The research provides the best accuracy so far in multidimensional stock signal forecast. A alternative evalution comes with Wavelet integration with sam structure and more. Interested persons can join the research. 2025-05-15 |
|
4. No Code Automation With LLM Models new
No code Automation : This project started with joint collaboration with founder of Stealth and T8 Innovation pvt Ltd in creation of a plugin in design automation. We have a few innovations in Cobra implementation that can disrupt the No code automation set up in LLM models. This is a problem adapted by multiple companies/startups who are currently working in design automation. A space of work that can go beyond usual LLMs. The work is progress.
Ultimate Imagination around the work : Given a scope of any application project, LLM needs to generate the code immediately in a specifc platform.
References related to the work : 2025-05-22 |
|
5. Financial ChatGPT new
Highest Imagination of the Work : A stromgest possible recommendation engine that can immediately generate a portfolio of the stock of companies where an arbitrary individual investor can choose to invest across all listed companies in all stock exchanges. The best recommendation engine for High frequency trading situation across all trading platforms/exchanges. ==================================================== Part of the Project : Ongoing Research Multimodal integration and collaboration opportunities: Extension of forescasting setup of Some variation of COBRA in sequential learning setup in large scale multi-modal learning with text and images. Interest students can follow the regular update of the page. You can write back in case you are in the line of the project.
2025-05-22 |
|
6. News on Alumni PhD Students new
Congratulations to Biplab Paul for joining Department of Mathematics, IIT Delhi as Assistant Professor 2025-04-01 |
|
7. Medication on Brain Haemorrhage due to Viral and other diseases new
Research focuses on Brain's adaptation on EEG signal and signal orientation on Brain and disease removal. Can core of the Human Brain Code change through injection of different sound frequencies ? Negative and positve sound evolutions and more. 2025-05-22 |
|
8. Conformal Prediction new
Signal Confidence interval Prediction is an important problem and taken up recently by a group of faculties around the world. Updates of some of our research results are available at (1) (2)
References : (1) (2) (3) 2025-05-25 View |
|
9. Our Innovation On Boosting Page new
Ping Cobra started from 2020 now available with regular update at \url{https://sites.google.com/view/pingcobra/home} 2025-05-25 |
