The analysis and evaluation method for small samples from the perspective of regression
YIN Boya1, LI Chenyi2
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12C) 1298-1301
1 Department of Mathematics college of science, Zhejiang University of Technology;
2 JianXing Honors College, Zhejiang Universities of Technology, Hangzhou City, Zhejiang Province, China, 310023
It is well-known that regression analysis often suffers from the small sample problem while big data is a necessity to greatly improve the credibility of research. We propose an analysis and evaluation method to judge the quality of small samples accurately through taking full advantage of these representative data. To achieve it, regression analysis was adopted to describe and expand the small amount of data. Meanwhile, principal component analysis would be performed to give the comprehensive evaluation of the forwarded data. We demonstrate the method by taking micro alloying of adding aluminium and rare earth as an example to explain the feasibility and accuracy of this analysis system. It proceeds as follows: stepwise regression data expansion and forwarding principal component analysis. The paper ends with recommendations that adding rare earth benefited the ensemble more to illustrate this analysis method can describe the general trend of data.