The Construction Of FMACD Index Based On SVM Multi-step Rolling Prediction

Recently,with the futures market price fluctuations more frequent,the traditional indicator of technical analysis methods has been unable to meet the needs of today's investors for market analysis and forecast.Traditional indicators have wide practicality,but it is easy resulting in the lag effect for the market price trend only by a simple calculation based on historical data to describe the occurred stock market.Because of the lag of general indicators,firstly we use SVR model to proceed multiple-step rolling forecast and obtain the highly accurate forecast price data.Then we combine forecast data with historical data by use the calculation method of MACD index in order to construct the FMACD indicators with the character of advance.This paper takes the daily main continuous data of SHFE CUM as an example,and gives the profit and loss of the MACD indicators and the FMACD indicators,respectively,as well as analyzes the reasons for the difference of simulating trading results,that is,the lag of MACD indicators may result in false judgement and delay,and thereby cause the loss,but the FMACD index has a character of trading signals in advance,being able to send out signals to enter in and out in time,which are more likely to benefit.Based on simulation trading of more than one thousand data and we find it takes only 0.154668 seconds by calculating the FMACD indicators,which is 30% higher than the traditional MACD indicators.So,the FMACD indicators can be applied to the simulation of 1 minute high-frequency trading.In the sequel,we take the SHFE CU1604 as an example to proceed a 1-minute high-frequency trade simulation,the results show that we can obtain more stable earnings by using FMACD indicators comparing to the MACD index,which demonstrate that FMACD indicators can also be applied to multi period situation,and highlight its extensively practical features.The construction method of FMACD indicators can be spreaded in the traditional index system by combining the predictive data with the historical data,which can make the traditional indicators advance and so it has more extensive application prospects.