| |
 |
UNIVERSITY OF BUCHAREST
FACULTY OF PHYSICS Guest
2024-11-22 2:29 |
 |
|
|
|
Conference: Bucharest University Faculty of Physics 2010 Meeting
Section: Atmosphere and Earth Science; Environment Protection
Title:
A comparative study of two types of expert system
Authors:
(1,2)Gina Tiron, (3)Steluţa Gosav, (4)Carolina-Irina Oprea, (2)Dana Ortansa Dorohoi2
Affiliation:
(1)Meteorological Centre Moldova, Theodor Vascauteanu St.10, Iasi, Romania,
(2)Al.I.Cuza University, Faculty of Physics, 11 Carol I Bldv. RO-700506, Iasi, Romania
(3)Physics Department, Faculty of Sciences, Dunărea de Jos University of Galati, 800201, Romania,
(4)Romatsa Bucuresti, Romania
E-mail
tiron.gina@.yahoo.com
Keywords:
artificial neural network,reflectivity,principal component analysis
Abstract:
In this paper we present two types of pure expert system: an artificial neural network model (ANN) and a hybrid artificial neural network model coupled with principal component analysis (PC-ANN.This ANN system uses as input reflectivity data from the WSR-98 D S-band Doppler radar located in Barnova, north-east Romania and some physical parameters from Aladin model i.e. wet bulb echipotential temperature, absolute vorticity, geopotential, humidity specific, air temperature, wind speed.This PC-ANN system uses as inputs the scores of the principal component obtained by the original data i.e. reflectivity intensity data and Aladin model parameters.This data are from 18-19 August 2005, 5 September 2007 and 24-27 July 2008.The validation results of all expert systems are analyzed.
|
|
|
|