Non-parametric Modeling and Its Application on Real-life Data

URI http://harp.lib.hiroshima-u.ac.jp/it-hiroshima/metadata/10646
File
Title
Non-parametric Modeling and Its Application on Real-life Data
Author
氏名 Pal Prasenjit
ヨミ
別名
氏名 Pal Subhabaha
ヨミ
別名
氏名 Pal Satyabrata
ヨミ
別名
氏名 Kageyama Sanpei
ヨミ カゲヤマ サンペイ
別名 景山 三平
Subject
Kernel
Spline
Loess
Parametric model
Non-parametric model
R2 criterion
Abstract

Some non-parametric models (kernel, loess and spline) have been compared with different parametric models in order to judge their precision by measuring the proximity between the observed and the predicted values based on the data-sets on the pest incidence (measured in terms of the number of pests) in respect of two important crops, namely, brinjal and chilli, with a view to describing the pattern of the incidence of pests over the year based on weekly incidence data (28 weeks). The importance of such studies stems from the fact that the farmers need the specialized scientific information in the form of advisory services in respect of the pattern of pest incidence over the year in advance and creation of pest-models helps us to obtain such important information. This paper is devoted to generate pest incidence models in cases of brinjal and chilli crops. Such pieces of generated information assist the farmers in their planning and execution of the cultivation of the said crops over the year. It is revealed that the precision levels (as are obtained by calculating the actual values of the well-known R2 criterion) are much higher when the non-parametric models are applied on the data-sets considered under the purview of the paper, in comparison to other methods exploited to obtaining the above-said information.

Journal Title
広島工業大学紀要. 研究編
Volume
46
Spage
173
Epage
179
Published Date
2012-02
Publisher
広島工業大学
ISSN
1346-9975
NCID
AA11599110
NAID
40019182685
Language
eng
NIIType
Departmental Bulletin Paper
Text Version
出版社版
Old URI
Set
it-hiroshima