Author : Christopher J. Beu
Publisher :
ISBN 13 : 9781369795684
Total Pages : pages
Book Rating : 4.7/5 (956 download)
Book Synopsis Non-metric Partial Least Squares Regression by : Christopher J. Beu
Download or read book Non-metric Partial Least Squares Regression written by Christopher J. Beu and published by . This book was released on 2017 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: Recent technological developments in process engineering and an ever present need for high-quality and highly-efficient processes have resulted in a surge in statistical data analysis research. The research on these topics is highly diverse. One underrepresented area is the analysis of non-metric data – a topic of particular interest to industries such as pharmaceuticals, paper and pulp, and petrochemicals along with various non-engineering disciplines such as biology and psychology. The dominating psyche is to avoid the variables that do not conform to the demands of the metric analysis. Here within, a thorough examination of two approaches to the statistical analysis of non-metric variables, binary partial least squares regression and optimal scaling partial least squares regression are explored. This analysis is then placed alongside the prevailing method – ignoring said variables. The results strongly support the inclusion of non-metric variables in capturing underlying process information. Percent variance explained is increased by 2 to 30 percent in each case study. However, the two previously mentioned methods do not provide such a stark conclusion. Ultimately, consideration of the variables is necessary but both non-metric analysis methods illuminate different traits in the variables. Furthermore, both methods used together underscore qualities of the variables that were previously absent.