[Publications | Committees | Patents | Contact]

    pmelvil@us.ibm.com

I am a Research Scientist in the Predictive Modeling Group at IBM T.J. Watson Research Center. My research interests lie in Machine Learning and Data Mining; more specifically, I am interested in active learning, ensemble methods, active feature-value acquisition, sentiment analysis, dual supervision, recommender systems, class probability estimation, semi-supervised learning, text classification, and applications of data mining to analyzing blogs, business analytics and e-commerce.  Our IBM Research team won the KDD Cup 2008 and the INFORMS Data Mining Contest 2008.

I got my Ph.D. from the Department of Computer Sciences at the University of Texas at Austin. At UT, I was a member of the Machine Learning group, led by Raymond Mooney. Prior to that I got bachelor's degrees in Computer Science and in Math from Brandeis University - where I worked at the DEMO Lab, Interaction Lab, and the Vision Lab.

Publications

  • Sentiment Analysis of Blogs by Combining Lexical Knowledge with Text Classification. [PDF]
    Prem Melville, Wojciech Gryc, and Richard Lawrence.
    In Proceedings of the 15th Conference on Knowledge Discovery and Data Mining (KDD-09), Paris, France, June, 2009.

  • Uncertainty Sampling and Transductive Experimental Design for Active Dual Supervision. [PDF]
    Vikas Sindhwani  and Prem Melville.
    In Proceedings of the 26th International Conference on Machine Learning (ICML-09), Montreal, Canada, June, 2009.

  • Active Dual Supervision: Reducing the Cost of Annotating Examples and Features. [PDF]
    Prem Melville and Vikas Sindhwani.
    In Proceedings of the NAACL HLT 2009 Workshop on Active Learning for Natural Language Processing, Boulder, Colorado, June, 2009.

  • Data Quality from Crowdsourcing: A Study of Annotation Selection Criteria. [PDF]
    Pei-Yun Hsueh, Prem Melville and Vikas Sindhawni.
    In Proceedings of the NAACL HLT 2009 Workshop on Active Learning for Natural Language Processing, Boulder, Colorado, June, 2009.

  • Active Feature-value Acquisition. [PDF]
    Maytal Saar-Tsechansky, Prem Melville, and Foster Provost.
    Management Science, Vol. 55, No. 4, pp. 664–684, April 2009.

  • Prediction-time Active Feature-value Acquisition for Customer Targeting. [PDF]
    Pallika Kanani and Prem Melville.
    In Proceedings of  the NIPS 2008 Workshop on Cost Sensitive Learning, Whistler, Canada, December 12, 2008.

  • Document-Word Co-Regularization for Semi-supervised Sentiment Analysis. [PDF][Extended Version]
    Vikas Sindhwani and Prem Melville.
    In Proceedings of IEEE International Conference on Data Mining (ICDM-08), Pisa, Italy, December 15-19, 2008.

  • Customer Targeting Models Using Actively-Selected Web Content. [PDF]
    Prem Melville, Saharon Rosset, and Richard Lawrence.
    In Proceedings of 14th Conference on Knowledge Discovery and Data Mining(KDD-08), Las Vegas, August 24-27, 2008.

  • Using Predictive Analysis to Improve Invoice-to-Cash Collection. [PDF]
    Sai Zeng, Prem Melville, Christian Lang, Ioana Boier-Martin, and Conrad Murphy.
    In Proceedings of 14th Conference on Knowledge Discovery and Data Mining(KDD-08), Las Vegas, August 24-27, 2008.

  • Breast Cancer Identification: KDD CUP Winner's Report. [PDF]
    Claudia Perlich, Prem Melville, Yan Liu, Grzegorz Swirszcz, Richard Lawrence, Saharon Rosset.
    SIGKDD Explorations, Vol. 10, Issue 2, 39-42, 2008.

  • Winner’s Report: KDD CUP Breast Cancer Identification. [PDF]
    Claudia Perlich, Prem Melville, Yan Liu, Saharon Rosset, Grzegorz Swirszcz, and Richard Lawrence.
    In Proceedings of the KDD-08 Workshop on Mining Medical Data, Las Vegas, August 24-27, 2008. (Workshop version of the above.)

  • Integrating Data Modeling and Dynamic Optimization using Constrained
    Reinforcement Learning.     [PDF]
    Naoki Abe, Prem Melville, Chandan K. Reddy, Cezar Pendus and David L. Jensen.
    Technical Report, 2008.

  • Learning Blog Sentiment with Reduced Supervision. [PDF]
    Prem Melville, Yan Liu, Wojciech Gryc, Richard Lawrence, Claudia Perlich.
    Technical Report, 2008.

  • Cost-Effective Clustering through Active Feature-value Acquisition. [PDF]
    Duy Vu, Prem Melville, Mikhail Bilenko, and Maytal Saar-Tsechansky.
    Technical Report, 2008.

  • A Machine-Learning Approach to Discovering Company Home Pages. [PDF]
    Wojciech Gryc, Prem Melville, and Richard Lawrence.
    Technical Report, 2008.

  • Data Acquisition and Cost-Effective Predictive Modeling: Targeting Offers for Electronic Commerce. [PDF]
    Foster Provost, Prem Melville, and Maytal Saar-Tsechansky.
    In Proceedings of the Ninth International Conference on Electronic Commerce, 2007.

  • Finding New Customers using Unstructured and Structured Data. [PDF]
    Prem Melville, Yan Liu, Richard Lawrence, Ildar Khabibrakhmanov, Cezar Pendus, and Timothy Bowden.
    In Proceedings of the KDD-07 Workshop on Mining Multiple Information Sources, 2007.

  • Intelligent Information Acquisition for Improved Clustering. [PDF]
    Duy Vu, Prem Melville, Mikhail Bilenko, and Maytal Saar-Tsechansky.
    In Workshop on Information Technologies and Systems (WITS), 2007.

  • Predictive Modeling for Collections of Accounts Receivable. [PDF]
    Sai Zeng, Prem Melville, Christian Lang, Ioana Boier-Martin, and Conrad Murphy.
    In Proceedings of the KDD-07 Workshop Domain Driven Data Mining, 2007.

  • An Expected Utility Approach to Active Feature-value Acquisition. [PDF]
    Prem Melville, Maytal Saar-Tsechansky, Foster Provost, and Raymond Mooney.
    In Proceedings of the Fifth IEEE International Conference on Data Mining (ICDM-05), 2005.

  • Economical Active Feature-value Acquisition through Expected Utility Estimation[Abstract] [Gzipped PS] [PDF]
    Melville, P., Saar-Tsechansky, M., Provost, F. and Mooney, R.J.
    Proceedings of the KDD-05 Workshop on Utility-Based Data Mining, Chicago, IL, August 2005.

  • Active Learning for Probability Estimation using Jensen-Shannon Divergence [Abstract] [Gzipped PS] [PDF]
    Prem Melville, Stewart M. Yang, Maytal Saar-Tsechansky, and Raymond J. Mooney
    In Proceedings of The 16th European Conference on Machine Learning (ECML), Porto, Portugal, 2005.

  • Combining Bias and Variance Reduction Techniques for Regression [Abstract] [Gzipped PS] [PDF]
    Yuk Lai Suen, Prem Melville and Raymond J. Mooney
    In Proceedings of The 16th European Conference on Machine Learning (ECML), Porto, Portugal, 2005

  • Creating Diverse Ensemble Classifiers to Reduce Supervision. [PDF]
    Prem Melville
    Ph.D. Thesis, Department of Computer Sciences, University of Texas at Austin, November 2005.
    Also appears as Technical Report TR-05-49, Artificial Intelligence Lab, University of Texas at Austin, December 2005.

  • Active Feature Acquisition for Classifier Induction [Abstract] [Gzipped PS] [PDF] [Tech Report]
    Prem Melville, Maytal Saar-Tsechansky, Foster Provost, and Raymond J. Mooney.
    In Proceedings of the Fourth International Conference on Data Mining (ICDM-2004). Brighton, UK. November 2004.
    Also appears as Technical Report UT-AI-TR-04-311, Artificial Intelligence Lab, University of Texas at Austin, Feb 2004.

  • Diverse Ensembles for Active Learning [Abstract] [Gzipped PS] [PDF][Tech Report]
    Prem Melville and Raymond J. Mooney.
    Proceedings of the 21st International Conference on Machine Learning     (ICML-2004), pp. 584-591, Banff, Canada, July 2004.

  • Experiments on Ensembles with Missing and Noisy Data[Abstract] [Gzipped PS] [PDF]
    Prem Melville, Nishit Shah, Lilyana Mihalkova, and Raymond J. Mooney
    Proceedings of the Fifth International Workshop on Multiple Classifier Systems (MCS-2004), F. Roli, J. Kittler, and T. Windeatt (Eds.), Lecture Notes in Computer Science, Vol. 3077, pp. 293-302, Cagliari, Italy, Springer Verlag, June 2004.

  • Creating Diversity in Ensembles Using Artificial Data [Abstract] [Gzipped PS] [PDF]
    Prem Melville and Raymond J. Mooney
    Information Fusion: Special Issue on Diversity in Multiclassifier Systems, 2004.

  • Relational Data Mining with Inductive Logic Programming for Link Discovery [Abstract] [Gzipped PS] [PDF]
    Raymond J. Moonney, Prem Melville, Lappoon R. Tang, Jude Shavlik, Inês de Castro Dutra, and David Page
    Data Mining: Next Generation Challenges and Future Directions, H. Kargupta, A. Joshi, K. Srivakumar, and Y. Yesha (Eds.), AAAI Press, 2004.

  • Creating Diverse Ensemble Classifiers [Abstract] [Gzipped PS] [PDF]
    Prem Melville
    Ph.D. proposal, Department of Computer Sciences, University of Texas at Austin, Oct 2003.
    Also appears as Technical Report UT-AI-TR-03-306, Artificial Intelligence Lab, University of Texas at Austin, December 2003.

  • Constructing Diverse Classifier Ensembles Using Artificial Training Examples [Abstract] [Gzipped PS] [PDF]
    Prem Melville and Raymond J. Mooney
    Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence (IJCAI-03), pp. 505-510, Acapulco, Mexico, August, 2003.

  • Scaling up ILP to Large Examples: Results on Link Discovery for Counter-Terrorism [Abstract] [Gzipped PS] [PDF]
    Lappoon R. Tang, Raymond J. Mooney and Prem Melville
    Proceedings of the KDD-2003 Workshop on Multi-Relational Data Mining (MRDM-2003), Washington DC, August 2003.

  • Relational Data Mining with Inductive Logic Programming for Link Discovery [Abstract] [Gzipped PS] [PDF]
    Raymond J. Mooney, Prem Melville, Lappoon R. Tang, Jude Shavlik, Inês de Castro Dutra, David Page and Vítor Santos Costa
    Proceedings of the National Science Foundation Workshop on Next Generation Data Mining, Baltimore, MD, November 2002.
    (Workshop version of the book chapter.)

  • Content-Boosted Collaborative Filtering for Improved Recommendations [Abstract] [Gzipped PS] [PDF]
    Prem Melville, Raymond J. Mooney and Ramadass Nagarajan
    Proceedings of the Eighteenth National Conference on Artificial Intelligence (AAAI-2002), pp. 187-192, Edmonton, Canada, July 2002.

  • Content-Boosted Collaborative Filtering [Abstract] [Gzipped PS] [PDF]
    Prem Melville, Raymond J. Mooney and Ramadass Nagarajan
    Proceedings of the SIGIR-2001 Workshop on Recommender Systems, New Orleans, LA, September 2001. (Workshop version of the above.)

  • Natural language assistant: A Dialog System for Online Product Recommendation [PDF]
    J. Chai, V. Horvath, N. Nicolov, M. Stys, N. Kambhatla, W. Zadrozny, and P. Melville
    AI Magazine    , 23(2):63--76, 2002.


Program Committees

  • 15th Conference on Knowledge Discovery and Data Mining, Paris, France, June 28- July 1, 2009 (KDD 2009).
  • 21st International Joint Conference on Artificial Intelligence, Pasadena, California, July 11-17, 2009 (IJCAI 2009).
  • 10th ACM Conference on Electronic Commerce, Stanford, California, July 6-10, 2009 (EC 2009).
  • European Conference on Machine Learning, Bled, Slovenia, September 7-11, 2009 (ECML 2009).
  • 18th ACM Conference on Information and Knowledge Management, Hong Kong, China, November 2-6, 2009 (CIKM 2009).
  • NAACL 2009 Workshop on Active Learning for NLP, Boulder, Colorado, June 5, 2009.
  • 23rd Conference on Artificial Intelligence, Chicago, IL, July 13-17, 2008 (AAAI 2008).
  • IEEE International Conference on Data Mining, Pisa, Italy, December 15-19, 2008 (ICDM 2008).
  • European Conference on Machine Learning, Antwerp, Belgium, September 15-19, 2008 (ECML 2008). 
  • 17th Conference on Information and Knowledge Management, Napa Valley, California
    October 26-30, 2008 (CIKM 2008).
  • NIPS Workshop on Cost-Sensitive Learning,  Vancouver, Canada, December 13th, 2008.
  • KDD Workshop on Mining Multiple Information Sources, Las Vegas, Nevada, August 24, 2008 (MMIS 2008).
  • Thirteenth Conference on Knowledge Discovery and Data Mining, San Jose, August 12-15, 2007 (KDD 2007).
  • 22nd Conference on Artificial Intelligence, Vancouver, Canada, July 22-26, 2007 (AAAI 2007).
  • Eighteenth European Conference on Machine Learning, Warsaw, Poland, September 17-21, 2007 (ECML 2007).
  • Twelfth Conference on Knowledge Discovery and Data Mining, Philadelphia, PA, August 20-23, 2006 (KDD 2006).
  • 23rd International Conference on Machine Learning, Pittsburgh, PA, June 25-29, 2006 (ICML 2006).
  • 21st National Conference on Artificial Intelligence, Boston, MA, July 16-20, 2006, (AAAI 2006).
  • Seventeenth European Conference on Machine Learning, Berlin, Germany, September 18-22, 2006 (ECML 2006).
  • Second KDD Workshop on Utility-Based Data Mining, Philadelphia, PA, August 20, 2006 (UBDM 2006).
  • Sixteenth European Conference on Machine Learning, Porto, Portugal, October 3-7, 2005, (ECML 2005).
  • First KDD Workshop on Utility-Based Data Mining, Chicago, IL, August 21, 2005 (UBDM 2005).

Patents

  • "Method and system for debt collection optimization", Naoki Abe, Prem Melville, Richard Lawrence, Cezar Pendus, David Jensen, Vince Thomas, Jim Bennett, and Tim Gardinier, United States Patent 7,519,553, April 14, 2009.
  • "A system and method for integrated graph-based learning using alternative sources of information", Vikas Sindhwani, Prem Melville, Yan Liu, and Richard Lawrence, filed 2008.   
  • "A method and system for identifying companies with specific business objectives", Prem Melville, Richard Lawrence, Yan Liu, Ildar Khabibrakhmanov, Upendra Chitnis, Timothy Bowden, filed 2008. 
  • "Method and system using machine learning to automatically discover company homes pages on the Internet", Prem Melville, Wojciech Gryc, Richard Lawrence, Upendra D. Chitnis, Ildar Khabibrakhmanov, and Cezar Pendus, filed 2008. 

 

Contact

IBM T.J. Watson Research Center

1101 Kitchawan Rd, Route 134/PO Box 218

Yorktown Heights, NY 10598

pmelvil@us.ibm.com