Research Interests

  • Semantic Web
  • Data Integration
  • Data Analytics
  • Data Mining
  • Machine Learning

Recent and Ongoing Research Projects

  • Semantic Web-based Geospatial Data Integration: Developing a semantic geospatial data integration framework, that addresses several types of heterogeneities in geospatial data, while simultaneously capturing uncertainty in various phases of data integration.
  • GUIDES – Geospatial Urban Infrastructure Data Engineering Solutions: Geospatial data management framework for urban underground infrastructure systems.
  • Predictive Analytics for Malaria Elimination in Zimbabwe: A predictive analytics framework to identify the malaria incidence using various indicators that affect the cause and spread of the disease at various spatio-temporal resolutions.
  • Ontology-based Instance Matching: A framework to integrate Food Inspections and Business Licenses dataset for the City of Chicago.
  • Ontology-based Data Exploration: Spatio-temporal analysis of Crime data using ontologies for the City of Chicago.
  • Ontology Alignment Evaluation Initiative (OAEI): Development of algorithms for spatial data matching and instance matching using AgreementMakerLight (AML). Participated in OAEI 2015, 2016, 2017 and 2017.5.
  • Automatic Extraction of Ontologies for Earth Science Ontology Repository and Ontology Matching for Linkipedia: An Entity Linking Tool, which takes semi- or un-structured text, extracts terms and link them to entities in a knowledge base
  • Making a Robust and Useful Earth Science Ontology Repository: Creating a test suite, automating ontology uploading, and standardizing RESTful API.

Other Academic Projects (Selected)

  • User Involvement in Ontology Matching: A semi-automatic ontology matching system using a hybrid active learning and online learning approach.
  • Geospatial Ontology Development: A mediator ontology to access data from tables, and is rich enough to encompass whichever domain ontologies are needed for the table collections to be managed.
  • Algorithms for Ontology Matching using AgreementMaker: Matching several domain-specific ontologies and use reference alignments as a benchmark to compute the corresponding alignments while significantly improving the overall F-measure.
  • Digit classification using Neural Networks: Hand-written digit recognition using Multiclass Perceptron Training Algorithm.
  • Traffic Sign Recognition System: Used Deep Convolutional Neural Networks for the detection of traffic signs and an ELM based approach for the classification task.
  • Sentiment Analysis & Opinion Mining for Tweet Classification: For the tweets collected from the 2012 US Presidential election, sentiment analysis was used to infer the polls trend by capturing the individual’s opinions against each Presidential candidate (Barack Obama and Mitt Romney).
  • LaBTReS - A Language Based Tag Recommender System: Recommendation of tags for Internet users who use languages except English, based on the user profile and geolocation.
  • Research and Analysis on an Optimal Solution for Hiding Sensitive Frequent Item Sets: Algorithms based on the notion of hybrid database generation to protect sensitive knowledge from being mined.