Please use this identifier to cite or link to this item:
Title:  Reduction of number of parameters and forecasting convective developments at Kolkata (22.53ºN, 88.33ºE), India during pre-monsoon season: An application of multivariate techniques
Authors: Chatterjee, Subhra
Ghosh, S
De, U K
Keywords: Linear discriminant analysis (LDA)
Principal component analysis (PCA)
Convective development (CD)
Fair weather (FW)
Issue Date: Oct-2009
Publisher: CSIR
Abstract: The present work aims at the formation of a statistical index using the most effective parameters out of primarily selected 20 parameters to predict the convective development at Kolkata (22.53ºN, 88.33ºE), India during the pre-monsoon season along with a multivariate statistical analysis of 20 different thermodynamic and dynamic parameters. In the first phase, linear discriminant functions formed with the help of linear discriminant analysis (LDA) of 20 parameters derived from the radiosonde observations at 0000 and 1200 hrs GMT during 1985-1996 at Kolkata produced 42.8 and 72.7% correct prediction for fair weather and convective development in next 12 hrs for the years 1997 to 1999 whereas in the evening the respective correct results are 50.8 and 56.6%. In the second phase, an attempt has been made to reduce the number of parameters using eigenvector analysis and correlation study between 20 original parameters and significant principal components. Here, again LDA has been performed with the parameters thus selected. The study reveals that for morning and afternoon, the number of parameters can be reduced to 6 and 7, respectively. The accuracy of predictions is also found to improve as compared to the first phase. However, it may be mentioned that total number of days involved in the present study is much less than the total number of days in the entire pre-monsoon season from 1985 to 1999 due to the unavailability of basic data needed to compute the required parameters
Description: 275-282
ISSN: 0975-105X (Online); 0367-8393 (Print)
Appears in Collections:IJRSP Vol.38(5) [October 2009]

Files in This Item:
File Description SizeFormat 
IJRSP 38(5) 275-282.pdf123.67 kBAdobe PDFView/Open

Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.