From 2013 to present,

His research focuses on learning representations for unstructured data especially focused on spatio-temporal datasets and machine learning with multiple instances. He is currently working on forecasting with many variables, understanding relations between multiple predictors and multiple-instance learning. Published papers can be reached through publications section.

Accepted research proposals:

• AFOSR (Air Force Office of Scientific Research, USA) Program: Research Interests of the Air Force Office of Scientific Research (Start date: 01.02.2017, End date 01.02.2020) 
  Modeling Unevenly Spaced Multivariate Time Series with Mixed Variable Types
• BAP (Boğaziçi University Scientific Research Projects) Program: SUP (Start-UP) (Start date: 01.11.2014, End date 01.11.2017) 
  Enhanced ensemble mechanisms for time series data mining
• TÜBİTAK (The Scientific and Technological Research Council of Turkey) Program: 2232 (Start date: 01.03.2014, End date 01.03.2016) 
  Finding robust representations for financial time series modeling using machine learning approaches

From 2010 to 2013,

• Big Data in Large Communication Networks: Mining and Visualization:  Passive monitoring of the communication networks requires storage of the captured traffic either in packet capture  or flow format. Network generates terabyte scale data during the capture and efficient methods are required to process and analysis of the data to characterize and understand the traffic. He is working on data reduction, visualization and analysis tools to discover the underlying relations in near-real time.
• Finding Efficient Representations for Time Series and Image Classification: Finding an efficient representation of high dimensional data to classification is an important mathematical problem. For images, invariance to global transformations such as translations, rotations or scaling is required. Time series classification approaches also suffer from similar problems. He is mainly working on local feature extraction approaches. Using supervised and unsupervised learning approaches, his focus is on finding efficient representations for the classification of the time series and images. The summary of the work is provided in Time Series Data Mining link.
• Evaluating the Impact of Interactive Tutoring and Game-Based Environments on Learning and Engagement: He is working on developing tools to manage and analyze the large sets of sensory data. This data mainly includes the multiple sources of information collected in a learning environment. More information about the project is available at http://angle.lab.asu.edu/site/.

From 2009 to 2010, he worked on:

• Multi-scale, Multi-feature Process Informatics for Mining Multimodal Sensory Data and Making Decisions in Sensorimotor Tasks: He processed neuronal data, worked on features to extract characteristics of the behavior of the neurons under different experiment conditions in this work.
• Models of Quality of Service and Quality of Information Assurance towards Their Dynamic Adaptation: He mainly worked on mathematical models for QoS requirements of service-based software systems, analyzed the QoS tradeoffs. He developed a GA for for real time decision of service parameters.

From 2007 to 2008, He worked on his M.S. Thesis titled:

Energy efficient coverage and connectivity problem in wireless sensor networks: Location of wireless sensors and routing of the data generated to a base station under two conflicting objectives: minimization of network cost and maximization of network lifetime, developed mathematical models and multiobjective genetic algorithm to solve the problem