Font Size: a A A

Shanghai Screw Steel Futures Price Prediction Based On BP Neural Network Model

Posted on:2013-11-04Degree:MasterType:Thesis
Country:ChinaCandidate:K J ChenFull Text:PDF
GTID:2268330425959811Subject:Software engineering
Abstract/Summary:PDF Full Text Request
The important position of the steel in the national economy is self-evident, it is irreplaceable and important supplies of nation-building. For a long time, due to the impact of the upstream iron ore price shocking and the dramatic changing of consumption of downstream enterprises, steel price fluctuates violently, and the majority of the iron and steel enterprises, steel traders and steel manufacturers are facing a huge price risk. Steel futures market is an extension of the spot market, with the risk diversification and price discovery function.This feature that makes more and more iron and steel enterprises, steel traders to participate in the futures market to lock-in costs or profits, reduce the price risk to a minimum.In the futures market, hedgers is important, they are the basis for the futures market that can be generated and the presence of; Risk diversification and price discovery function of futures market has realized through a lot of investors from investing activities, they constitute the microscopic foundation of the futures market.Precisely because of this, in this article,we have studied the deformed bar futures from the perspective of futures investment, based on BP neural network model,and at the two levels from the long-period (days) and short-period (5minutes) to predict the price of it.And we also analyized the earnings performance and the ability to predict price trends of the model.From the experiment results of the BP neural network modcl,the prediction accuracy of the short-period deformed bar futures price performs better than the long-period,and according to the performance of effect of the judgment of trends for the future price,the long period is better than the short-period. The experimental results prompt us:in the short term, especially in the high-frequency environment, the more the regularity of price changes; in the long term, the specific point of the price is bit difficult to determine, but the price shows a stronger trend.From the point of profit performance, all models in this article are verygood. But it must be noted that the profitability of the model is built on the condition of zero transaction costs, which is also the direction of our future efforts.
Keywords/Search Tags:Deformed bar, Futures Price, Back-Propagation Neural Network, Pridiction
PDF Full Text Request
Related items