Multi-layer Perceptron and Pruning
Journal Title: Turkish Journal of Forecasting - Year 2017, Vol 1, Issue 1
Abstract
A Multi-Layer Perceptron (MLP) defines a family of artificial neural networks often used in TS modeling and forecasting. Because of its “black box” aspect, many researchers refuse to use it. Moreover, the optimization (often based on the exhaustive approach where “all” configurations are tested) and learning phases of this artificial intelligence tool (often based on the Levenberg-Marquardt algorithm; LMA) are weaknesses of this approach (exhaustively and local minima). These two tasks must be repeated depending on the knowledge of each new problem studied, making the process, long, laborious and not systematically robust. In this short communication, a pruning process is presented. This method allows, during the training phase, to carry out an inputs selecting method activating (or not) inter-nodes connections in order to verify if forecasting is improved. We propose to use iteratively the popular damped least-squares method to activate inputs and neurons. A first pass is applied to 10% of the learning sample to determine weights significantly different from 0 and delete other. Then a classical batch process based on LMA is used with the new MLP. The validation is done using 25 measured meteorological TS and crosscomparing the prediction results of the classical LMA and the 2-stage LMA.
Authors and Affiliations
Cyril Voyant, Christophe Paoli, Marie-Laure Nivet, Gilles Notton, Alexis Fouilloy, Fabrice Motte
Keywords
Related Articles
- EP ID EP299759
- DOI -
- Views 133
- Downloads 0
A Comparison of Support Vector Regression and Multivariable Grey Model for Short-Term Wind Speed Forecasting
Wind energy is one of the most promising resources of energy for the future. Wind is generally regarded as the most renewable and green energy type. The reason for this perception is mainly because of wind’s inexhaustibl...
G-STAR Model for Forecasting Space-Time Variation of Temperature in Northern Ethiopia
Among many indicators of climate change, the temperature is a key indicator to take remedial action for world global warming. This finding provides application of space-time models for temperature data, which is selected...
Forecasting of Apricot Production of Turkey by Using Box-Jenkins Method
Turkey is the first largest apricot producer in the world. In 2016, Turkey was responsible for 9,21% of world apricot production with 730 thousand tons. Turkey also generated 11,31% of world apricot exports in 2016. The...
Time Series Prediction with Direct and Recurrent Neural Networks
This article presents a comparative study of the prediction of time series for the Consumer Price Index (CPI) using a recurrent neural network (RNN). For this, three models are designed for recurrent networks, with chang...
Why and How Does Exponential Smoothing Fail? An In Depth Comparison of ATA-Simple and Simple Exponential Smoothing.
Even though exponential smoothing (ES) is publicized as one of the most successful forecasting methods in the time series literature and it is widely used in practice due to its simplicity, its accuracy can be affected b...