Optimization ofWavelet Neural Networks Model by Setting theWeighted Value of Output Through Fuzzy Rules Takagi-Sugeno-Kang (TSK) Type As a Fixed Parameter

Syamsul Bahri, Widodo and Subanar (2018) Optimization ofWavelet Neural Networks Model by Setting theWeighted Value of Output Through Fuzzy Rules Takagi-Sugeno-Kang (TSK) Type As a Fixed Parameter.

[img]
Preview
Text
Artikel Journal Internasional-GJPAM-2016.pdf

Download (1MB) | Preview

Abstract

In this paper, we propose a model of fuzzy wavelet neural network (FWNN) which is optimized from wavelet neural network (WNN) model by setting the weighting coefficient of the output using the TSK fuzzy inference type. This coefficient in the proposed FWNN model is seen as exogenous parameters that are not updated in the learning process using the method of gradient descent with momentum. The accuracy and the execution time of the model was illustrated using several univariate time series data cases, and the results were compared with models ofWNNthat had been previously published. The simulation results for variety of these cases show that the effectiveness and the accuracy of proposed FWNN model is better than the previous model of WNN.

Item Type: Article
Keywords (Kata Kunci): Wavelet neural network, fuzzy wavelet, wavelet B-spline, fuzzy inference, TSK fuzzy rules, time series.
Divisions: Fakultas Matematika dan ilmu Pengetahuan Alam
Depositing User: Muslimin Muslimin
Date Deposited: 14 Nov 2018 03:31
Last Modified: 14 Nov 2018 03:31
URI: http://eprints.unram.ac.id/id/eprint/10090

Actions (login required)

View Item View Item