Comparative study among different neural net learning algorithms applied to rainfall time series

Chattopadhyay, Surajit; Chattopadhyay, Goutami

doi:10.1002/met.71

Cited by 47 publications

(25 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Selvam et al, 1995;Sivakumar, 2001;Dix and Hunt, 2007;Chattopadhyay and Chattopadhyay, 2008a) and the suitability of an ANN in dealing with chaotic time series is well established (Prinicipe et al, 1992). In the present study, the ANN approach has been adopted to forecast precipitation in a time step ahead.…”

Section: Introductionmentioning

confidence: 98%

“…Hung et al (2008) used meteorological parameters in developing an ANN rainfall forecast models. Chattopadhyay (2007) applied an ANN in the form of multilayer perceptron to forecast monsoon rainfall over India using some meteorological predictors and Chattopadhyay and Chattopadhyay (2008a) compared the performances of various backpropagation learning algorithms in forecasting monsoon rainfall time series using the monthly mean rainfall amounts as predictors. Chattopadhyay and Chattopadhyay (2008b) compared the performances of different ANNs with variable hidden layer size in forecasting rainfall time series.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Artificial neural network models for forecasting intermittent monthly precipitation in arid regions

Dahamsheh

Aksoy

2009

Meteorological Applications

View full text Add to dashboard Cite

Forecasting monthly precipitation in arid regions is investigated by means of feed forward back propagation (FFBP) artificial neural networks (ANNs) and compared to the linear regression technique with multiple inputs (MLR). Four meteorological stations from different geographical regions in Jordan are selected. The ANNs and MLR processes are analysed based on the mean square error, relative/absolute error, determination coefficient as well as the central statistical moments such as mean, standard deviation, and minimum and maximum values. It is found that whilst on one hand the ANNs are slightly better than the MLR in forecasting the monthly total precipitation, on the other hand, both are found with to have limitations which should be improved by means of either changing the type and architecture of the ANNs or incorporating modelling tools such as Markov chains into the forecast model.

show abstract

Section: Introductionmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

Artificial neural network models for forecasting intermittent monthly precipitation in arid regions

Dahamsheh

Aksoy

2009

Meteorological Applications

View full text Add to dashboard Cite

show abstract

“…It should be mentioned that there is no strict rule to decide the ratio of training and test cases. A survey of the ANN literature found that the ratios 1 : 1 (Chattopadhyay and Chattopadhyay, 2008a), 7 : 3 (Lundin et al, 1999) and 3 : 1 (Perez et al, 2000) are frequently used in ANN applications. In the present paper an approach similar to that of Lundin et al (1999) has been adopted after examining the other approaches.…”

Section: Development Of Artificial Neural Network Modelmentioning

confidence: 99%

“…Willmott's indices of order 1 and 2, which are highly acceptable measures of goodness of fit (Willmott, 1982;Chattopadhyay and Chattopadhyay, 2008a) in meteorological modelling, are given in Table III. It is revealed that the maximum values of Willmott's indices of order 1 and 2 are produced in June and the values 0.657 (for order 1) and 0.873 (for order 2) are considerably high in this case.…”

Section: The Measure Of Error Of Estimation (Ee) Is Expressedmentioning

confidence: 99%

Estimating potential evapotranspiration from limited weather data over Gangetic West Bengal, India: a neurocomputing approach

Chattopadhyay¹,

Jain²,

Chattopadhyay

2009

Meteorological Applications

Self Cite

View full text Add to dashboard Cite

This paper reports the development of a neurocomputing-based model for estimating the potential evapotranspiration over Gangetic West Bengal, India during the summer monsoon months of June, July and August. An artificial neural network is implemented in the form of multilayer perceptron to generate the model. Three weather variables, surface temperature, vapour pressure and rainfall are used as the independent variables in generating the model. The performance of the model is judged statistically against non-linear regression in the form of asymptotic regression. The study reveals that an artificial neural network is more efficient than the regression approach to estimate the potential evapotranspiration in the summer monsoon months. Furthermore, it is established that the artificial neural network and non-linear regression have almost equal efficiency in the previously mentioned estimation in the month of June. However, in July and August the higher values of correlation and Willmott's indices, and lower values of estimation error, indicate that the artificial neural network is more reliable than the non-linear regression approach. Since evapotranspiration is one of the basic components of the hydrological cycle and is essential for estimating irrigation water requirement, an efficient estimation procedure may help in agrometeorological modelling and irrigation scheduling in the summer monsoon months, which are of high importance for agriculture in the study zone.

show abstract

“…Their model results 0.8 correlations between actual and predicted rainfall. Surajit and Goutami Chattopadhyay (2008) identify a non linear methodology to forecast the time series of average summer monsoon rainfall over India [8]. The RMSE for four NN based techniques is 0.15, 0.26, 0.42 and 0.47.…”

Section: Introductionmentioning

confidence: 99%

A Fast Learning Algorithm for Rainfall Prediction

Mahajan¹,

Mazumdar²

2014

IJCA

View full text Add to dashboard Cite

A PC based application is developed using 51 years of Indian rainfall data for long range prediction of average rain fall. This learning algorithm iteratively estimates 96 coefficients of a 5th order polynomial in few minutes. Proposed prediction model is based on modelling of time series rainfall data using 5th order non-linear predicting code. Steepest descent algorithm is used for extraction of appropriate coefficient of rainfall time series data. These coefficients are reinforced in model learning process. The rainfall data of 1960 to 2010 is used for the development of the model. Model has been tested on rainfall data for different training sets. The proposed model is capable of forecasting yearly rainfall 1 year in advance. Rainfall estimation accuracy is above 85%.

show abstract

Comparative study among different neural net learning algorithms applied to rainfall time series

Cited by 47 publications

References 35 publications

Artificial neural network models for forecasting intermittent monthly precipitation in arid regions

Artificial neural network models for forecasting intermittent monthly precipitation in arid regions

Estimating potential evapotranspiration from limited weather data over Gangetic West Bengal, India: a neurocomputing approach

A Fast Learning Algorithm for Rainfall Prediction

Contact Info

Product

Resources

About