Start Submission Become a Reviewer

Reading: REDUCTION OF EXPERIMENTAL ERROR IN COCONUT WITH ADJUSTMENT BY AN INTEGRATED INDEX DEVELOPED ...

Download

A- A+
dyslexia friendly

Articles

REDUCTION OF EXPERIMENTAL ERROR IN COCONUT WITH ADJUSTMENT BY AN INTEGRATED INDEX DEVELOPED THROUGH PRINCIPAL COMPONENT ANALYSIS USING VEGETATIVE AND REPRODUCTIVE CHARACTERS

Authors:

D. T. MATHES ,

COCOS, LK
About D. T.
Coconut Research Institute, Lunuwila, Sri Lanka.
X close

H. A. J. GUNATHILAKA,

LK
About H. A. J.
Coconut Research Institute, Lunuwila, Sri Lanka.
X close

D. N. S. FERNANDO

LK
About D. N. S.
Coconut Research Institute, Lunuwila, Sri Lanka.
X close

Abstract

Precision in estimation and accuracy of the results from experiments is dependent
on the nature of experimental variation. Controlling variation through calibration,.
other than through local control, is a well defined concept since 1953.
It is important that the choice of a calibrating variate or variates to be, cost
effective, less time consumsing in recording, and efficient in reducing the
experimental error. For experiments in coconut, one and two years pre-experimental
yield is used as calibrating variates. Collecting yield data over a period of one or two
 years to use as a calibrating variate is a long wait before commencement of an
experiment: Financially too it is high.
* This paper explores the merits of using, an 'Integrated Index' (incorporating the
vegetative and reproductive characters) as a calibrating variate evaluated through the
technique of 'Principal Component Analysis'. The characters were, .'Trunk Girth
(X,)', 'Number of Inflorescences (Xj)', and Number of nuts per bunch (X3)', and
'Number of green fronds QQ\ The Integrated Index - I' developed through
principal component analysis reflects the 'Vigour of the Pahn'.
The eigen vectors of the four characters and the eigen values of the four vectors
were evaluated through Principal Component Analysis. The standardized, vector
scores for each palm was taken as the Integrated Index for each palm. The average
of these values for a given experimental plot was taken as the Index for each.plot
and was used as the calibrating variate.
The results suggest that, the overall range in the values for the coefficient of
variation' when adjusted for T was 7.05 to 10.78% as compared to the range, 8.35
to 21.02% observed without adjustment.
The recording of variables X, to X« was completed within 3 to 4 days before the
53
commencement of the trial. Thus recording of the characters was less time
consuming and cost effective.
How to Cite: MATHES, D.T., GUNATHILAKA, H.A.J. & FERNANDO, D.N.S., (2010). REDUCTION OF EXPERIMENTAL ERROR IN COCONUT WITH ADJUSTMENT BY AN INTEGRATED INDEX DEVELOPED THROUGH PRINCIPAL COMPONENT ANALYSIS USING VEGETATIVE AND REPRODUCTIVE CHARACTERS. COCOS. 11, pp.53–68. DOI: http://doi.org/10.4038/cocos.v11i0.2158
Published on 10 Aug 2010.
Peer Reviewed

Downloads

  • PDF (EN)

    comments powered by Disqus