|
Activity Number:
|
405
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #302497 |
|
Title:
|
Regression Analysis for Line Scale Data in Sensory Science
|
|
Author(s):
|
Lin Xie*+ and Rui Xiong+
|
|
Companies:
|
Kansas State University and Unilever Home & Personal Care
|
|
Address:
|
1541 International Ct, N12, Manhattan, KS, 66502, 20 Hull St, Consumer Science Insights, Shelton, CT, 66502,
|
|
Keywords:
|
Regression analysis ; line scale data ; preference mapping
|
|
Abstract:
|
Line scales are widely used in sensory evaluation and consumer testing. The scores of response variables from line scales take only values within a specified range (e.g., 0-10, 0-100). In preference mapping and sensory-instrumental relationship, the response variables are often linearly regressed on some other explanatory variables. Since regression models are inherently unbounded (-8, +8), they could produce predicted values that are outside the ranges of the line scales for some response variables. Conventionally, if predicted values are outside the specified range, they are set to either the minimum or maximum value of the scales. In the paper, an alternative approach, which transforms the line scale range to unbounded range (-8, +8) before regression analysis, is proposed to address the problem.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2008 program |
