Dissertation/Thesis Abstract

Prediction and classification of organoleptic quality of high density polyethylene resin by means of multivariate statistical analysis of volatile organic compound profiles
by Mathews, Eric B., M.S., California State University, Long Beach, 2015, 131; 1601317
Abstract (Summary)

High Density Polyethylene (HDPE) is a commonly used polymer for food and beverage packaging, but tends to have issues with plastic off-taste. Current industry practice for monitoring and controlling taste quality of HDPE packaging involves sensory analysis by a panel of human analysts. This method of data collection is time consuming, labor intensive, and subjective. As an alternative, multivariate statistical methods may be used to correlate volatile organic compound profiles of HDPE packaging to organoleptic quality of the material.

This study explores the use of multiple linear regressions, principal component regression, and partial least squares regression for predicting organoleptic quality of HDPE packaging. Cross validation was performed for all models to estimate prediction performance. Results of cross validation show that partial least squares in combination with variable selection lead to the best prediction performance. Lastly, linear discriminant analysis was performed for classifying HDPE packaging into organoleptic quality categories.

Indexing (document details)
Advisor: Sung, Kim E.
Commitee: Chaidez, Manuel, Moon, Hojin, Zhou, Tianni
School: California State University, Long Beach
Department: Mathematics and Statistics
School Location: United States -- California
Source: MAI 55/02M(E), Masters Abstracts International
Subjects: Statistics, Packaging
Keywords: Gas chromatography, High density polyethylene, Multivariate statistics, Off-taste, Patial least squares, Sensory
Publication Number: 1601317
ISBN: 9781339123899