Research on Grey Forecasting Model with Acceleration Characteristics

碩士 === 義守大學 === 工業工程與管理學系碩士班 === 97 === From the past to now, human being is curious about the unknown future. Subsequently, numerous methods are developed to predict the future, from “Bagua (The Eight Diagrams)” of ancient China and “Astrology” of Babylon to the “Tarot” which is most popular in the...

Full description

Bibliographic Details
Main Authors: Chih-hao Lin, 林志豪
Other Authors: Chun-i Chen
Format: Others
Language:zh-TW
Published: 2009
Online Access:http://ndltd.ncl.edu.tw/handle/74408483106238019927
Description
Summary:碩士 === 義守大學 === 工業工程與管理學系碩士班 === 97 === From the past to now, human being is curious about the unknown future. Subsequently, numerous methods are developed to predict the future, from “Bagua (The Eight Diagrams)” of ancient China and “Astrology” of Babylon to the “Tarot” which is most popular in the present time. All the efforts are devoted to predict the future and try to avoid the disaster and get the fortune. Due to the development of mathematics and statistics, the forecasting science is a fast growing subject. In the recent years, there is an emerging forecasting method which is called “Grey Theory” proposed by Prof. Deng. The characteristics of the Grey theory are easy mathematics, less data needed and high prediction precision. Among the Grey theory, the Grey forecasting is one of the most popular subjects. The general case is GM(m,n) which m represents the order of differentiation and n represents the number of influencing factors. The GM(1,1) is the basic model and is applied to numerous research fields. The GM(2,1) is the research interest in this study because of its controversy in existence. In order to show its existence, the interpretation of GM(2,1) is introduced by considering physical point of view. And it is believed that GM(2,1) indeed exists and is suitable for system with acceleration or deceleration because of second differentiation appeared in the original equation. To prove our deduction correct, a set of raw data with periodical change is chosen and collected. The international cruel oil price obtains the characteristic of periodical change and its future trend is forecast both by GM(1,1) and GM(2,1) respectively. The result shows that GM(2,1) holds higher forecasting accuracy than GM(1,1) because of the periodical change essence in cruel oil price. Futhermore, the GM(2,1) is applied to forecast the domestic oil price and the result is also satisfactory.