Release from the Curse of High Dimensional Data Analysis

Golub et al. started their research to find oncogenes and new cancer subclasses from microarray around 1970. They opened their microarray on the Internet. The other five medical projects published their papers and released their microarrays, also. However, because Japanese cancer specialist advised us that NIH decided that these researches were useless after 2004, we guess medical groups abandoned these researches. Although we are looking for NIH’s report, we cannot find it now. Meanwhile, many researchers of statistics, machine learning and bioengineering continue to research as a new theme of high-dimensional data analysis using microarrays. However, they could not succeed in cancer gene analysis as same as medical researches (Problem5). We discriminated six microarrays by Revised IP-OLDF (RIP) and solved Problem5 within 54 days until December 20, 2015. We obtained the two surprising results. First, MNMs of six microarrays are zero (Fact3). Second, RIP could decompose microarray into many linearly separable gene subspaces (SMs) and noise subspace (Fact4). These two new facts indicate that we are free from the curse of high dimensional microarray data and complete the cancer gene analysis. Because all SMs are LSD and small samples, we thought to analysis all SMs by statistical methods and obtained useful results. However, we were disappointed that statistical methods do not show linearly separable facts and are useless for cancer gene diagnosis (Problem6). After trial and error, we make signal data made by RIP discriminant scores (RipDSs) from SM. Through this breakthrough, we find useful information by correlation analysis, cluster analysis, and PCA in addition to RIP, Revised LP-OLDF and hard margin SVM (H-SVM). We think that the discovery of the above two new facts is the essence of Problem5. Moreover, we claim to solve Prpblem6 and obtain useful medical care information from signal data as a cancer gene diagnosis. However, our claim needs validation by medical specialists. In this research, we introduce the reason why no researchers could succeed in the cancer gene diagnosis by microarrays from 1970.

[1]  Ian B. Jeffery,et al.  Comparison and evaluation of methods for generating differentially expressed gene lists from microarray data , 2006, BMC Bioinformatics.

[2]  R. Gentleman,et al.  Gene expression profile of adult T-cell acute lymphocytic leukemia identifies distinct subsets of patients with different response to therapy and survival. , 2004, Blood.

[3]  Venkatesan Guruswami,et al.  Combinatorial feature selection problems , 2000, Proceedings 41st Annual Symposium on Foundations of Computer Science.

[4]  E. Lander,et al.  Gene expression correlates of clinical prostate cancer behavior. , 2002, Cancer cell.

[5]  M. R. Mickey,et al.  Estimation of Error Rates in Discriminant Analysis , 1968 .

[6]  Claudio De Stefano,et al.  An Experimental Comparison of Feature-Selection and Classification Methods for Microarray Datasets , 2019, Inf..

[7]  U. Alon,et al.  Broad patterns of gene expression revealed by clustering analysis of tumor and normal colon tissues probed by oligonucleotide arrays. , 1999, Proceedings of the National Academy of Sciences of the United States of America.

[8]  D. Firth Bias reduction of maximum likelihood estimates , 1993 .

[9]  Shuichi Shinmura Problems of Discriminant Analysis by Mark Sense Test Data , 2011 .

[10]  Guoqing Diao,et al.  Assessing Genome-Wide Statistical Significance for Large p Small n Problems , 2013, Genetics.

[11]  Shuichi Shinmura,et al.  New Theory of Discriminant Analysis and Cancer Gene Analysis , 2019, High-dimensional Microarray Data Analysis.

[12]  J. Friedman Regularized Discriminant Analysis , 1989 .

[13]  Shuichi Shinmura End of Discriminant Functions based on Variance-covariance Matrices , 2014, ICORES.

[14]  Antonie Stam,et al.  Nontraditional approaches to statistical classification: Some perspectives on L_p-norm methods , 1997, Ann. Oper. Res..

[15]  M. Aoshima,et al.  Distance-based classifier by data transformation for high-dimension, strongly spiked eigenvalue models , 2017, 1710.10768.

[16]  Shuichi Shinmura Cancer Gene Analysis of Microarray Data , 2018, 2018 IEEE International Conference on Big Data, Cloud Computing, Data Science & Engineering (BCD).

[17]  Shuichi Shinmura Four Serious Problems and New Facts of the Discriminant Analysis , 2014, ICORES.

[18]  Matthias Dehmer,et al.  Analysis of Microarray Data , 2008 .

[19]  H. Riedwyl,et al.  Multivariate Statistics: A Practical Approach , 1988 .

[20]  Todd,et al.  Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning , 2002, Nature Medicine.

[21]  Bing Chen,et al.  User-Personalized Review Rating Prediction Method Based on Review Text Content and User-Item Rating Matrix , 2018, Inf..

[22]  Sara van de Geer,et al.  Statistics for High-Dimensional Data: Methods, Theory and Applications , 2011 .

[23]  D. Cox The Regression Analysis of Binary Sequences , 1958 .

[24]  Shuichi Shinmura New Theory of Discriminant Analysis After R. Fisher , 2016 .

[25]  M. Aoshima,et al.  High-Dimensional Quadratic Classifiers in Non-sparse Settings , 2015, Methodology and Computing in Applied Probability.

[26]  J. Mesirov,et al.  Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. , 1999, Science.

[27]  F. Zhan,et al.  The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. , 2003, The New England journal of medicine.

[28]  S. Shinmura,et al.  A NEW ALGORITHM OF THE LINEAR DISCRIMINANT FUNCTION USING INTEGER PROGRAMMING , 2000 .

[29]  新村 秀一 Optimal Linear Discriminant Functions using Mathematical Programming , 2000 .

[30]  Sara van de Geer,et al.  Statistics for High-Dimensional Data , 2011 .