My research interest

1. Data Mining
2. Support Vector Machine
3. Boosting
4. Neural Networks
5. Machine Learning
6. Statistical Physics
   Self Organized Criticality
7. Web Text Mining

My publication list

1. Statistical Mechanics
   
   

   • Ground state of one-dimensional chiral XY model under external magnetic field and uniaxial potential.
     M. Momma and T. Horiguchi, Physica A 234 (1997) 837.

   • Finite temperature properties of one-dimensional chiral XY model under an external field and a uniaxial potential.
     M. Momma and T. Horiguchi, Physica A, 251 (1998) 485.

   • Ground state of antiferromagnetic thin films of XY model under magnetic field and uniaxial potential.
     M. Momma and T. Horiguchi, Physica A, 259 (1998) 105.

2. Neural Nets
3. Barker-Fock Model
   
   

   • Multiple reentrance and fractal phase boundaries in the barker-fock model.
     M. Momma and J.S. Walker, Physica A, 278 (2000) 52.

   • One more paper on thermodynamic properties of the Barker-fock model. Will be submitting soon.

   
   
4. Support Vector Machines and Machine Learning
   
   

   • A Pattern Search Method for Model Selection of Support Vector Regression.
     M. Momma and K.P. Bennett, Proceedings of SIAM Conference on Data Mining (2002).

   • MARK: A Boosting Algorithm for Heterogeneous Kernel Models.
     K.P. Bennett, M. Momma and M. Embrechts, Proceedings of SIGKDD International Conference on Knowledge Discovery and Data Mining (2002).

   • Sparse Kernel Partial Least Squares Regression.
     M. Momma and K.P. Bennett, Proceedings of Conference on Learning Theory, 2003.

   • Efficiently Learning the Metric using Side-Information.
     Tijl De Bie, Michinari Momma and Nello Cristianini, in Proc. of the 14th International Conference on Algorithmic Learning Theory (ALT2003), Sapporo, Japan, October 2003.

   • Constructing Orthogonal Latent Features
     Michinari Momma and Kristin Bennett, Feature Extraction, Foundations and Applications, Isabelle Guyon, Steve Gunn, Masoud Nikravesh, and Lofti Zadeh, editors, Springer 2005

   • Efficient Computations via Scalable Sparse Kernel Partial Least Squares and Boosted Latent Features
     Michinari Momma, Proceedings of SIGKDD International Conference on Knowledge Discovery and Data Mining (2005). (longer version)

About Me

	Japanese. 1995 Bachelor of Engineering in Electrical Communications Engineering, Tohoku University in Japan. 1997 Master of Information Sciences, Graduate School of Information Sciences, Tohoku University. 1999 Master of Science in Physics, Washington State University. Ph.D. in DSES (2003). My advisors were Prof. Bennett and Prof. Embrechts.

Link

Machine Learning, Collaborators

• Kristin Bennett
• Mark Embrechts
• Tijl De Bie
• Nello Cristianini

Statistical Physics

• The Physical Society of Japan
• A.-L. Barabasi
• Bak-Sneppen model

Japanese

Please send your comments to me!