Articles

Found 2 Documents
Search

System for Prediction of Non Stationary Time Series based on the Wavelet Radial Bases Function Neural Network Model Kusdarwati, Heni; Handoyo, Samingun
International Journal of Electrical and Computer Engineering (IJECE) Vol 8, No 4: August 2018
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Original Source | Check in Google Scholar

Abstract

This paper proposes and examines the performance of a hybrid model called the wavelet radial bases function neural networks (WRBFNN). The model will be compared its performance with the wavelet feed forward neural networks (WFFN model by developing a prediction or forecasting system that considers two types of input formats: input9 and input17, and also considers 4 types of non-stationary time series data. The MODWT transform is used to generate wavelet and smooth coefficients, in which several elements of both coefficients are chosen in a particular way to serve as inputs to the NN model in both RBFNN and FFNN models. The performance of both WRBFNN and WFFNN models is evaluated by using MAPE and MSE value indicators, while the computation process of the two models is compared using two indicators, many epoch, and length of training. In stationary benchmark data, all models have a performance with very high accuracy. The WRBFNN9 model is the most superior model in nonstationary data containing linear trend elements, while the WFFNN17 model performs best on non-stationary data with the non-linear trend and seasonal elements. In terms of speed in computing, the WRBFNN model is superior with a much smaller number of epochs and much shorter training time.
Penerapan Bagan Kendali Multivariat Robust Pada Data Produksi Pupuk ZK PT Petrokimia Gresik Darmanto, Darmanto; Kusdarwati, Heni; Iriany, Atiek; Setiawan, Iwan; Ashari, Ayu Aisyah
Performa: Media Ilmiah Teknik Industri Vol 17, No 1 (2018): PERFORMA Vol. 17, No 1 Maret 2018
Publisher : Program Studi Teknik Industri, Universitas Sebelas Maret

Show Abstract | Original Source | Check in Google Scholar | Full PDF (306.19 KB)

Abstract

PT Petrokimia Gresik is the most complete fertilizer producer in Indonesia and one of its production is ZK fertilizer. There are five measurable chemicals that correlate to form ZK fertilizer ie H2O, H2SO4, K2O, SO3 and Cl-. ZK fertilizer monitoring process has not been statistically done by PT Petrokimia Gresik, either univariat or multivariate. Since ZK fertilizer is composed of five chemicals that correlate each other, a multivariate control chart is used. RMCD is one of the robust parameter estimation methods for outlier data. The average vector and variance-covariance matrix derived from the RMCD method is used to calculate the statistics on the multivariate control chart. Therefore, the robust control chart is more sensitive to detecting a shift in production processes compared to the classical ones. The data used in Phase I is daily data per January 1 - April 30, 2017, while Phase II data used is daily data as of May 1 - July 15, 2017. The results of the control chart analysis in Phase I shows that the production process has not been controlled statistically analysis of cause-effect diagrams. Furthermore, the control chart limits in Phase I that have been stable after the repair are used for Phase II production data. The result of the control chart analysis in Phase II shows that the production process has shifted. This can be known by the number of points that out of control.