Homepage of
      G. Praveen Kumar

142 Halsey Dr, Apt 7
West Lafayette, Indiana - 47906
Phone : +1 765 637 1367
Email : gpraveenkumar5[at]ymail[dot]com
Homepage : http://gpraveenkumar.com
If we knew what it was we were doing,
it would not be called research, would it?
-- Albert Einstein

Get in Touch :

   Biography       Research Interests       Research Experience       Work Experience       Publications       Projects       Courses Undertaken       Achievements       Position of Responsibility       Blog

More Projects

  • Exact inference for a positive Markov random field
  • March 13
    • Implemented maximum cardinality variable elimination ordering to get maximal cliques in a graph. Constructed a clique tree using a maximum spanning tree algorithm. Performed calibration using belief propagation to compute the univariate marginal probabilities for all variables. Given evidence, estimated posterior probability for the rest of univariate posterior marginals.
  • Categorization of Yelp Reviews
  • Sep 12 – Dec 12
    • Built Naive Bayes classifiers on Yelp's Academic dataset to classify reviews as Cool, Useful, Funny, Positive and generated learning curves. Applied smoothing methods and feature selection (using unigram vs. bigram features). Used K-Means to clustered data with latitude, longitude, review count as features.
  • Unparser and Interporcedural Analysis on GCC-4.7.0
  • Sep 12 – Dec 12
    • Built an unparser that traverses the abstract syntax tree to regenerates the C source code. Performed dataflow analysis on function call graphs to identify the use of uninitialized global and local variables.
  • Movie Recommendation System
  • April 12
    • Implemented Memory-Based Collaborative Filtering Algorithm to make recommendations for movies. Also, developed new methods to increase the recommendation accuracy and improve the overall F1 measure.
  • Text Categorization for Detecting Email Spam
  • Mar 12
    • Developed a simple text categorization system for detecting email spam using the Naive Bayes text categorization algorithm and studied the empirical results with a real world email collection. Compared features from relevant (i.e., spam) and in the irrelevant models. Applied smoothing methods and feature selection to improve the categorization performance.
  • Software Security
  • Nov 11 -- Dec 11
    • Performed a series of software vulnerability exploits (carried out Buffer Overflow, Format String and return to libc attacks), explored unsafe and insecure programming techniques (like dangers of executable stack) and evaluated the efficacy of operating system defences against them.
  • MiniDB Database
  • Sep 11 – Dec 11
    • Built in three parts, namely Buffer Manager, Relational Operator and Query Evaluation, a working single-user relational database management system. Started from scratch with a few basic components and used Java programming to implement the functions and various interfaces that involved several hundred lines of code.
  • Web Security
  • Sep 11 – Oct 11
    • Carried out a series of attacks like cross site scripting, cross site request forgery, and SQL injection attacks to exploit some vulnerabilities in an online bulletin-board. Implemented modifications to the web site to prevent such attacks. Also performed a brute-force dictionary attack to recover passwords of 100 users using most frequent password list obtained from the Web.
  • Online Admission System
  • Jan 10 – April 10
    • Lead a team of 7 members as a course project for Software Engineering Lab to automate the admission system, pre-check the inclusion of all required material and automatically rank each student’s application based on a number of criteria. Developed a website by incorporating necessary logic, with JSP as the front end technology, Websphere and Oracle Database servers as the backend.
  • An Intelligent Library
  • Jan 10 – Feb10
    • Incorporated the concepts of Data Mining like Weighted Association Rule Mining and Clustering to develop an efficient recommender system for Library and thereby reducing both the search process and time.
  • MyLibrary
  • Sep 09 – Jan 10
    • Made a website for the scenario of Online Library Management, for a competition conducted by IBM “The Great Mind Challenge 2009”. AJAX and JSP were used for designing the front end. Used IBM technologies like DB2 for the database and WebSphere as a server for the the back end.
      Project Guide: Mr. Jaydeep Howlader
      Assistant Professor, Department of Information Technology, NIT Durgapur
  • Similarity Framework
  • May 09 - July 09
    • Developed a Framework in Java that integrates Perl scripts for the computation of similar tags in Folksonomy or like data, for the Knowledge Data Engineering Group, University of Kassel, Germany. Log4j was used in the Framework to capture logs.
      Project Guide: Dr. Gerd Stumme and Dr. Andreas Hotho
      Professor and Senior Researcher, Department of EE/CS, University of Kassel, Germany
  • Intelligent Cache management for Squid Proxy Server
  • Mar 09 - Oct 09
    • While caching is not traditionally a function of proxy servers, an increase in performance can be achieved by caching the contents of an accessed location. Subsequent requests for access will hence lead to already cached contents being used. Hence, we were involved in research to develop an effective mechanism for caching in proxy server. A project for Ekalavya, IIT Bombay.
      Project Guide: Mr. Parag Kumar Guhathakurtha
      Assistant Professor, Department of Computer Science and Engineering, NIT Durgapur
  • Sphynx Bank Online
  • Aug 08 - Dec 08
    • Built a website as an online comprehensive solution to Internet Banking for “The Great Mind Challenge 2008”, a competition conducted by IBM. Used IBM technologies like DB2 and WebSphere for the back end and JSP was used for designing the front end pages.
      Project Guide: Mr. Jaydeep Howlader
      Assistant Professor, Department of Information Technology, NIT Durgapur
  • The Ultimate Exam Simulator
  • May 08 - July 08
    • Modelled and simulated an online examination portal in-accordance to the futuristic needs of the Entrance Examination using core Java and MSSQL 7.0 for “SHARE 2008”, the summer project competition organized by Sun Microsystems.
  • Calculation of Torsional Strain
  • Mar 07 – Apr 07
    • A project done to calculate the torsional, normal strain and their diagrammatic representation in C++ for Department of Civil Engineering, NIT Durgapur.
      Project Guide: Dr. A.K. Dutta
      Professor, Department of Civil Engineering , NIT Durgapur
  • Cazino – The world’s number one tennis club
  • Aug 04 – Feb 06
    • A multiplayer tennis like game developed in C++ for Standard XII (high school) project. This entertainment based project uses graphics for liveliness and also maintains details about the club members.


    © 2011-2015 G. Praveen Kumar