Knowledge enhanced connectionist models for short-term electric load forecasting

Rahman, Saifur; Drezga, I.; Rajagopalan, J.

doi:10.1109/ann.1993.264314

Cited by 4 publications

(2 citation statements)

References 13 publications

(1 reference statement)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where r is an overlap parameter. After the parameters of the membership functions have been found, the weights in layer 4 are obtained by using the competitive learning algorithm (Rumelhart & Zipses, 1985) as follows:…”

Section: Phase-i: Unsupervised Learning Phasementioning

confidence: 99%

“…Because the input variables and the estimated ones from the hybrid neural network have wide variations in magnitudes, they will cause convergence problem and the system will behave completely erratic. To circumvent this problem, the variables are scaled between 0.1 and 0.9 (Rahman, Drezga, & Rajagopalan, 1993). This is performed so as to maximise accuracy and minimise training time.…”

Section: Scaling Of the Variable Rangementioning

confidence: 99%

See 1 more Smart Citation

Building a Fuzzy Expert System for Electric Load Forecasting Using a Hybrid Neural Network

Dash

Liew

Rahman

et al. 1995

Expert Systems with Applications

View full text Add to dashboard Cite

This paper presents the development of a hybrid neural network to model a fuzzy expert system for time series forecasting of electricc load. The hybrid neural network is trained to develop fuzzy logic rules andjind optimal inputloutput membership values of load and weather parameters. A hybrid learning algorithm consisting of unsupervised and supervised learning phases is used for training the fuzzifred neural network. In the supervised learning phase, both back-propagation and linear Kalman jilter algorithms are used for the adjustment of weights and membership functions. Extensive tests have been performed on a 2-year utility data for the generation of peak and average loadprojiles in 24 h. 48 h. and 168 h ahead time frame during summer and winter seasons. From the simulation results, it is observed that the fuzzy expert system using the Kalman $lter-based algorithm gives faster convergence and more accurate prediction of a load time series.

show abstract

Section: Phase-i: Unsupervised Learning Phasementioning

confidence: 99%

Section: Scaling Of the Variable Rangementioning

confidence: 99%

Building a Fuzzy Expert System for Electric Load Forecasting Using a Hybrid Neural Network

Dash

Liew

Rahman

et al. 1995

Expert Systems with Applications

View full text Add to dashboard Cite

show abstract

Fuzzy neural networks for time-series forecasting of electric load

Dash

Ramakrishna

Liew³

et al. 1995

IEE Proc., Gener. Transm. Distrib.

View full text Add to dashboard Cite

Three computing models, based on the multilayer perceptron and capable of fuzzy classification of patterns, are presented. The first type of fuzzy neural network uses the membership values of the linguistic properties of the past load and weather parameters and the output of the network is defined as fuzzy-class-membership values of the forecast load. The backpropagation algorithm is used to train the network. The second and third types of fuzzy neural network are developed based on the fact that any fuzzy expert system can be represented in the form of a feedforward neural network. These two types of fuzzy-neural-network model can be trained to develop fuzzy-logic rules and find optimal input/output membership values. A hybrid learning algorithm consisting of unsupervised and supervised learning phases is used to train the two models. Extensive tests have been performed on two-years of utility data for generation of peak and average load profiles 24 hours and 168 hours ahead, and results for typical winter and summer months are given to confirm the effectiveness of the three models. IntroductionThe application of artificial-neural-network-(ANN) and fuzzy-logic-based decision-support systems to time-series forecasting has gained attention recently [7-121. ANNbased load forecasts give large errors when the weather profile changes very fast. Also, extremely slow training or even training failure occurs in many cases owing to dificulties in selecting proper structures of the neuralnetwork paradigm being used, and owing to the errors in associated parameters such as learning rates, activation functions etc. which are fundamental to any backpropagation neural network. On the other hand, the development of a fuzzy decision system (fuzzy expert system) for load forecasting requires detailed analysis of data and the fuzzy-rule base has to be developed heuristically for each season. The rules fixed in this way may not always yield the best forecast. The shortcomings of the neural-network paradigm can be partly remedied by the recognition of the fact that the learning speed and accuracy of an ANN may often be enhanced by utilising the knowledge of neural-network expertise in a specific application. This human knowledge can be encoded by fuzzy expert systems, which are integrated into the fuzzy neural network (FNN). The present work is aimed at achieving a robust load forecast with much improved accuracy using three different models of FNNs. For the neural network to be called a FNN, the signal and/or the weights should be fuzzified [14]. The first type of FNN, abbreviated FNN,, is based on the multilayer perceptron, using the backpropagation algorithm. The input vector consists of the membership values of linguistic properties of the past load and weather parameters and the output vector is defined in terms of fuzzy-class-membership values of the forecast load. The second and third types of FNN, abbreviated as FNN, and FNN,, are based on the argument that any fuzzy expert system employing one block of rules may ...

show abstract

Load forecasting using artificial intelligence techniques: a literature survey

Bansal

Pandey

2005

IJCAT

View full text Add to dashboard Cite

Knowledge enhanced connectionist models for short-term electric load forecasting

Cited by 4 publications

References 13 publications

Building a Fuzzy Expert System for Electric Load Forecasting Using a Hybrid Neural Network

Building a Fuzzy Expert System for Electric Load Forecasting Using a Hybrid Neural Network

Fuzzy neural networks for time-series forecasting of electric load

Load forecasting using artificial intelligence techniques: a literature survey

Contact Info

Product

Resources

About