Functional link neural network – artificial bee colony for time series temperature prediction

摘要

Higher Order Neural Networks (HONNs) have emerged as animportant tool for time series prediction and have been successfully applied inmany engineering and scientific problems. One of the models in HONNs is aFunctional Link Neural Network (FLNN) known to be conveniently used forfunction approximation and can be extended for pattern recognition with fasterconvergence rate and lesser computational load compared to ordinaryfeedforward network like the Multilayer Perceptron (MLP). In training theFLNN, the mostly used algorithm is the Backpropagation (BP) learningalgorithm. However, one of the crucial problems with BP learning algorithm isthat it can be easily gets trapped on local minima. This paper proposed analternative learning scheme for the FLNN to be applied on temperatureforecasting by using Artificial Bee Colony (ABC) optimization algorithm. TheABC adopted in this work is known to have good exploration and exploitationcapabilities in searching optimal weight especially in numerical optimizationproblems. The result of the prediction made by FLNN-ABC is compared withthe original FLNN architecture and toward the end we found that FLNN-ABCgives better result in predicting the next-day ahead prediction.