Analysis on PLSR in Contrast with PCR

Abhimanyu Kumar

Abstract: In this paper, we have described a very efficient tool for data mining, i. e. , predictive data mining techniques. Partial Least Square Regression, predictive technique is briefly discussed and is compared with Principal Component Regression, another predictive technique. We have implemented Partial Least Square Regression (PLSR) technique on the Gasoline dataset. From other researchers, we have taken the result of Principal Component Regression (PCR) technique and compared our result of PLSR with the PCRs result, and it is observed that the result of PLSR is much more significant than the PCR, in respect to R-square error (R-sq) criteria for model comparison. Also we have seen that the 3 components of PLSR gives better result as compared to 4 components of PCR.

Keywords: KDD- knowledge Discovery database, PDM Predictive Data Mining, PCR Principal Component Regression, PLSR Partial Least Square Regression, PCA Principal Component Analysis

How to Cite?: Abhimanyu Kumar, "Analysis on PLSR in Contrast with PCR", Volume 4 Issue 6, June 2015, International Journal of Science and Research (IJSR), Pages: 2369-2371, https://www.ijsr.net/getabstract.php?paperid=SUB155741, DOI: https://dx.doi.org/10.21275/SUB155741