Transaction Price-based Futures Analysis And Prediction

According to futures trade professional prognostic three principles: The market behavior forgives everything, price to turn into by trend, the history will repeat, putting the point that the information mining only in the strike price, by analyzing the timing data coming from the current agreement we get the possible strike price of the next day, at the same time, by mining the relating relation between the current agreement and the historical agreement, we forecast the whole direction of the next period of time.Analyzing the timing list of the current agreement by ARIAM arithmetic, on this base, forecasting the strike price of next trade day. Aim at the calculate way of ARIMA arithmetic from not steady to steady limiting the result, put forward the time list many class faintness turns processed method, becoming to the same contract come into being many different thin turn bargains record, analyzing and forecasting every track record by ARIMA arithmetic, then form several forecasting value, and get the finally forecasting value by averaging. Comparing with immediacy analyzing and forecasting precision degree increase clearly, and receive expected target.Mining the correlative deal note from historical agreement by gray relative arithmetic, take gray relative degree as reliability, take corresponding historical agreement as the consult price of currently forecast. Let multilevel blur data joining into original data set, compute the correlative degree at the same class, then arithmetic average the multilevel correlative coefficient which belong to the same agreement and get the last result. By this way we partly solve the problem which pay much attention to part so as to ignore integer, there is clearly improvement comparing with using directly the original data to compute, and effectively weed the deal record which the whole direction patency reverse but the correlative coefficient very big.