Fusion of Statistics, Neural-Net, Machine Learning

1.挑战生物信息学我们的研究组与癌症研究所的基因组中心合作，取得了成功的单核苷酸多态(SNP)分型项目。这篇论文已于2004年发表在《生物信息学》杂志上，部分工作已申请日本专利局。目前，SNP单倍型阻断的下一个项目前景看好，并将在不久的将来取得结果。在这次研究合作中，我们每月召开一次研究会议，其中蛋白质组数据的研究现在取得了快速进展。2003年11月，北卡罗来纳州立大学P.K.Sen教授作了一次关于生物信息学和非参数方法的讲座，由我们的研究小组组织。2.统计理论和Boost算法在这三年中，村田博士、Kanainori博士、Takenouhi博士和江口博士在周六组织了每月一次的研究会议，在会上，村田博士、Kanainori博士、Takenouhi博士和江口博士建立了统计学上的理解，并提出了几种替代AdaBoost的变种。独立成分分析和主成分分析第四届ICA国际会议于2003年在奈良会议中心举行，村田博士、南博士和江口博士担任组委会。这项关于自组织主成分分析的研究发表在《机器学习研究》杂志上。

1.Challenging bioinformaticsOur research group results to succeed the project for SNP (Single Nucleotide Polymorphism) typing with co-operative work with the genome center in Research Institute of Cancer Study. The paper has published in the journal of 'Bioinformatics' in 2004 and the portion of the work has applied to Japan Patent Office. Presently the next project of SNP haplotype blocking is promising and will attain the result in a near future.In this research collaboration we have a monthly research meeting, in which the investigation for proteome date is now made a rapid progress. In November, 2003 Professor P.K. Sen at North Carolina State University gave a lecture on bioinformatics and nonparametric method, which was organized by our research group.2.Statistical theory and Boost algorithmIn these three years a monthly research meeting has been organized on Saturday in which Dr Murata, Dr Kanainori, Dr Takenouchi and Dr Eguchi has establish statistical understanding and several variants alternative to AdaBoost. Their two papers has published in a journal of 'Neural Computation' in 2004.3.Independent component analysis and principal component analysisThe fourth International conference on ICA had held at Nara Conference center in 2003, in which Dr Murata, Dr Minami and Dr Eguchi worked as organizing committee. The work on self-organizing PCA has published in J.Machine Learning Research.