Applying INGHS-based ANN in Predicting the Price Fluctuation of Taiwan Steel Products

• Main Authors: Shu-Ting Chou, 周姝廷
• Other Authors: Chui-Yu Chiu
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/pqr3m4

碩士 === 國立臺北科技大學 === 工業工程與管理系所 === 105 === Being affected by worldwide market, steel industry becomes the foundation of economic and industrial development. It is always considered as the strength of national characterization. Steel industry is also an index to measure the indicators of industrial development of a country. Harmony Search (HS), Artificial Neural Network (ANN) and Back-Propagation Network (BPN) have been applied to many researches in prediction, clustering and real problems. It is also proved that the performance is excellent. Therefore, the purpose of this study is to establish a steel price fluctuation predicting system which combines Improved Novel Global Harmony Search (INGHS) with ANN, as the reference when making decisions of research and operating for Taiwan Steel Industry. This study makes a comparison of NGHS-ANN, BPN and INGHS-ANN. The research result shows that INGHS-ANN has the best performance in predicting the price fluctuation of steel products.