Forecasting Computer Products Sales by Integrating Ensemble Empirical Mode Decomposition and Extreme Learning Machine

Abstract

A hybrid forecasting model that integrates ensemble empirical model decomposition (EEMD), and extreme learning machine (ELM) for computer products sales is proposed. The EEMD is a new piece of signal processing technology. It is based on the local characteristic time scales of a signal and could decompose the complicated signal into intrinsic mode functions (IMFs). The ELM is a novel learning algorithm for single-hidden-layer feedforward networks. In our proposed approach, the initial task is to apply the EEMD method to decompose the original sales data into a number of IMFs. The hidden useful information of the original data could be discovered in those IMFs. The IMFs are then integrated with the ELM method to develop an effective forecasting model for computer products sales. Experimental results from three real computer products sales data, including hard disk, display card, and notebook, showed that the proposed hybrid sales forecasting method outperforms the four comparative models and is an effective alternative for forecasting sales of computer products.

A Novel CEEMD-Based EELM Ensemble Learning Paradigm for Crude Oil Price Forecasting

“…34 First, the CEEMD algorithm, the current extension of the competitive decomposition techniques of EMD and EEMD, is implemented to decompose the original data of crude oil price time series into several relatively independent IMFs and one residue. Second, the EELM tool, a fast and powerful forecasting tool compared with other AI models (e.g., ANN and SVR), is utilized to predict the extracted components independently.…”

“…In existing models, various data decomposition algorithms have been utilized, e.g., EMD, 8 EEMD 33 and wavelet decomposition 35 ; and EEMD has been repeatedly shown to be the most e®ective tool in the EMD family. 33,34 However, there exists a serious drawback in the EEMD algorithm: the large level of residue noise in the signal reconstruction stemming from the added white noises. Even though the residue noise can be mitigated as the ensemble number increases, it would be a time-consuming process.…”

“…Finally, these predicted values are fused into an ensemble result as the¯nal prediction by simple addition (ADD) ensemble method, 33 since the sum of real values of the decomposed components actually equals to the original data. This novel CEEMD-based EELM model is actually an extension on existing ensemble models, 8,33,34 and the main contribution to literature is to introduce the current algorithms of CEEMD and EELM and to test their e®ectiveness in formulating the powerful \decomposition and ensemble" learning paradigm.…”

Financial Time Series Forecasting Using Empirical Mode Decomposition and Support Vector Regression

“…Forecasting results were computed exclusively for the following testing sample. We note that in the literature, EMD is instead sometimes applied to the whole dataset, including the testing part (see, e.g., Chen et al 2012;Cheng and Wei 2014;Lin et al 2012;Lu and Shao 2012;Wang et al 2014;Yu et al 2008). The inclusion of (future) testing data in the forecasting methodology is clearly wrong, providing meaningless "forecasts".…”

Predictions of Industrial and Commercial Electricity Sales in Taiwan Using ARIMA and Artificial Neural Networks Techniques

“…The fundamental concept an integrated forecasting scheme is to capture various patterns in the data by taking advantage of each individual model's capability. The findings have been reported that the two-stage modeling is superior for improving the performance of single-stage modeling [17]- [23]. Since both ARIMA and ANN are appropriate for predicting the IECES, this study considers an integrated ARIMA-ANN as the two-stage model.…”