Open AccessARTIFICIAL NEURAL NETWORK METHOD FOR ESTIMATION OF MISSING DATA
Author(s) -
HARSHANAND K. GHUGE,
D. G. Regulwar
Publication year - 2012
Publication title -
international journal of advanced technology in civil engineering
Language(s) - English
Resource type - Journals
ISSN - 2231-5721
DOI - 10.47893/ijatce.2012.1039
Subject(s) - unavailability , missing data , artificial neural network , computer science , data mining , rain gauge , estimation , current (fluid) , hydrological modelling , artificial intelligence , machine learning , statistics , engineering , mathematics , geology , telecommunications , radar , electrical engineering , systems engineering , climatology
The availability of precipitation data plays important role for analysis of various systems required for design of water resources systems. The perfect measurements are not available always. The scientist/hydrologists come across the problem of missing data due to a variety of reasons. There may be various reasons of unavailability of data. Measurement of hydrologic variables (e.g. rainfall, stream flows, etc.) is prone to various instrumental/systematic, manual and random errors. In the current study, missing rainfall data is evaluated by using Artificial Neural Network Method. Historical precipitation data from 6 rain-gauge stations in the Maharashtra State, India, are used to train and test the ANN method and derive conclusions from the improvements in result given by ANN. Results suggest that ANN model can be work for estimation of missing data.