Hi! I am Monir.

Md Monir Hossain, Tianyu Zhang, Omid Ardakanian, "Identifying grey-box thermal models with Bayesian neural networks", Energy and Buildings, 238, p.110836.

This paper explores various techniques for establishing a suitable thermal model using time series data generated by smart thermostats.

  • Proposing a methodology based on Bayesian neural networks for identifying the RC model of a home equipped with a smart thermostat.
  • Comparing the predictive power of different RC models and showing that a 2R2C model yields lower accuracy than other low-order models for most homes in our dataset.
  • Showing that a grey-box model is more accurate than several black-box models, including time series and neural network models, in predicting the room temperature.
  • Showing that clustering allows for transferring a pre-trained representative model of that cluster to this home with and without adaptation.
  • Discussing how a model trained for one season can be transferred to another season.

Md Monir Hossain, Mark Sebestyen, Dhruv Mayank, Omid Ardakanian, Hamzeh Khazaei, "Large-scale Data-driven Segmentation of Banking Customers", 2020 IEEE International Conference on Big Data (Big Data), 2020, pp. 4392-4401.

This paper presents a novel big data analytics framework for creating explainable personas for retail and business banking customers.

  • Using raw-data-based time series clustering techniques on a large-scale real banking transaction data from approximately 60,000 retail and 90,000 business customers and obtaining representative patterns for those clusters.
  • Using hierarchical clustering to find and analyse anomalous customers over various time frames.
  • Investigating cluster stability through different time-snaps and different time horizons.
  • Extracting useful and interesting rules by combining the metadata (demographic and financial) and the cluster identifier of each customer.
  • Evaluating the performance of the automated framework developed for customer segmentation and show that it is scalable horizontally and vertically.

Md Monir Hossain, Tianyu Zhang, Omid Ardakanian, "Evaluating the Feasibility of Reusing Pre-trained Thermal Models in the Residential Sector", In Proceedings of the 1st ACM International Workshop on Urban Building Energy Sensing, Controls, Big Data Analysis, and Visualization (UrbSys), 2019.

The goal of this work is to build thermal model at scale.

  • Developing algorithm based on Bayesian Neural Network for building RC model of homes based on ecobee dataset.
  • Using clustering to select representative homes from clusters of home.
  • Using transfer learning for transferring models from representative homes to other homes of the cluster with little or no retraining

Md Monir Hossain, Nima Mahmoudi, Changyuan Lin, Hamzeh Khazaei, and Abram Hindle. "Executability of Python Snippets in Stack Overflow." arXiv preprint arXiv:1907.04908 (2019).

Online resources today contain an abundant amount of code snippets for documentation, collaboration, learning, and problem-solving purposes. Their executability in a ‘plug and play’ manner enables us to confirm their quality and use them directly in projects. But, in practice that is often not the case due to several requirements violations or incompleteness. The goal is to analyze these snippets.

  • Developing a scalable framework to investigate Python Code Block executability in Stack Overflow using SOTorrent.

Md Monir Hossain, Shuvro Barua, and M. A. Matin. "A pre-feasibility study for electrification in Nijhum Dwip using hybrid renewable technology." In 2015 International Conference on Electrical & Electronic Engineering (ICEEE), pp. 225-228. IEEE, 2015.

  • Investigating power situation, conducting a renewable resource survey & estimating real load in Nijhum Dwip (an island)
  • Finding out an optimized hybrid power system for Nijhum Dwip using HOMER Renewable Energy Software & RETScreen