The Use of Multivariate Analyses to Eliminate Redundant Variables when Evaluating Potatoes for Chip Quality
1989
Loiselle, F. | Tai, G. C. C. | Tarn, T. R. | Christie, B. R.
Selection for good chipping clones in potato breeding often takes place during the middle and late selection stages. This is due to the considerable resources needed for the assessment of several characteristics. Objective multivariate methods for the elimination of redundant variables were applied to eight standardized chipping quality traits recorded on advanced clones over 15 years at two locations in New Brunswick, Canada. The clones were evaluated of specific gravity and the following chipping scores: direct from the field in August; after 10 days storage at 13 °C; from 4 °C storage in February; from 4 °C storage in February with 1 week, 2 weeks, 3 weeks and 4 weeks reconditioning at 16 °C. Five elimination methods produced a similar deletion order. In every subset obtained, specific gravity and direct chipping from the field were always retained. Reconditioning treatments from one to four weeks after cold storage were shown to be the most redundant. The reduction in the number of traits from eight to four by every technique could recover more than 80 % of the original variation, showing that redundant variables were effectively discarded.
显示更多 [+] 显示较少 [-]