Direction and degree of the change in exchange rate are the foundations of foreign exchange risk management, and are one of the important index http://forum.effizienzgurus.de signed in the international financial market all the more. The fluctuation of valuta brings international trade and capital ck mens underwear circulation risk from both macroscopic and microcosmic levels. For making people master the change of rate of exchange in advance and evade risk brought by foreign exchange fluctuations in international moncler down jackets economic activities. It’s more and more urgent to predict direction and degree of the change in exchange rate in the future stated period. Hence, the forecast and the research in RMB exchange rate have great theoretical significance and practical value.Currently, the forecast methods of the exchange rate are Structured Variable Forecasting method and Prediction of Time Series method etc. On the selection of tools, ANN (cheap nfl jerseys artificial neural network) and Bayes Estimate are in common used. Prediction of Time Series method is just based on the history data generally, so it doesn’t take full advantage of related chi hair straightener experiential knowledge http://forums.alliedmods.net and related impact factors. ANN can resolve how to create a complex non-linear mapping from parametric space to solution space. It is provided with the ability of predicting one or several targets, so the ANN is a good method to work outs the problem. However, forecasting with simplex ANN has its proper weaknesses such as network structure, network astringency, network generalization, the problem on appending new sample and so on. If we can disassemble the whole parametric space of impact factors to several parametric subspaces, and design the respective http://www.host-my-post.com/forum for parametric space, the problem mentioned above can be resolved well. Therefore the thesis introduces Dempster – Shafer Evidence Theory (DS Therory) that is one important theory in Data Fusion and optimizes BPNN (Back Propagation Neural Network) by GA (Genetic Algorithm) and finally carries out the exchange rate forecast model based on GA-BPNN and DS Evidence Theory triumphantly according to the characteristics of the ANN and DS Evidence Theory.First, this thesis summarizes the research progress and theoretical achievements of the domestic and abroad exchange rate forecast technique, and it also elaborates various classical deciding theories of the exchange rate combing the characteristics of exchange rate as the models’ importing structural variable and sets up the Pretreatment level of the model.
Previous Entry: the macro economy
Next Entry: basic principles
Comments are closed, but trackbacks and pingbacks are open.