Research Of Commodity Index Which Based On Dynamic Adaptive Hedonic Model

With the rapid development of commodity markets and the increasing maturity of commodity index, commodity index plays a growing prominent role in commodity trading, and its variations have an impact on the stock market, bond market, and foreign exchange market. However, most of the current commodity indexes mainly focus on futures pricing, while the commodity price index for the spot market is still relatively few. When compile the stock index, we always have to face the problems arising from the non-standardized products and quality characteristics of different issues in the spot trading. Solving a series of challenging questions, such as how to incorporate the characteristics into modeling, how to determine the relationship between commodity prices and the characteristic attributes, and how to evaluate the"pure price change"of the commodities through the sample data, is a breakthrough in the research of the spot commodity index, as well as a expansion of the commodity index theories and methods.In this work, we study the main features of commodity market price based on the commodity price index theory and the hedonic model, establish the commodity index model, and provide practical theory tools for making government policy, deciding the supplier trading strategy, and determining the purchaser consumer decision. Finally we provide a systematic processing guidance for commodity spot index including 1) the early data analysis and selection;2) the design of index system;3) the selection of the commodity composition;4) the preparation of the model; 5) the pretreatment of the input data to the optimization; 6) empirical calculations in the end. In addition, we introduce the principal component analysis method into Hedonic Index Model in order to reduce the input dimensions for improving collinearity among the variables, relevance, and significance level. Consequently, the structure of Hedonic Index Model is simplified, and its convergence and effectiveness are improved.In this thesis, we firstly establish the framework of the commodity index system based on the general index design principles and ideas, determine the commodity composition on the basis of existing data, and confirm the weight distribution according to the flow volume index which is transactions volume. Secondly, we use the improved principle component analysis to analyze and adjust the characteristic attribution variables of the commodity trading information, besides improve and make adaptive rectification in the Hedonic index model. Then we input the principal components of the selected variables as the factor to the designed Hedonic Index Model and complete the establishment of the commodity index model. The empirical calculations analysis is performed on the trading data from a large commodity information service provider and the Shanghai Futures Exchange rubber trading. In the dynamic adaptive hedonic model, the transaction price per unit weight for the rubber is used as the dependent variable, while the variety features, the original production characteristics, and market characteristics are used as independent variables. We analyze and compare the empirical results. The results confirm the validity of the model and the improved effect.