Combining HHT signal separation and machine learning techniques for computer products sales forecasting
| Autor: | Lin, ChengHan, 林承翰 |
|---|---|
| Rok vydání: | 2011 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 99 Computer hi-tech products are the essential commodities in recent years. Technology techniques develop rapidly cause high products replacement rate. Business manager have to perform a good sales forecasting to solve the apply chain problem. This study integrates signal separation and machine learning techniques to construct sales forecasting model. To use empirical mode function (EMD) and ensemble empirical mode function (EEMD), two methods of Hilbert-Huang transform (HHT) to pre-process the sales data. Then, using two of machine learning techniques, support vector machine (SVR) and extreme learning machine (ELM) as forecasting tools. Constructing sales forecasting models and to compare. This study uses computer hi-tech products sales data to construct sales model by above-mention methods. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|