Study On Chaos Characteristic Of Soybean Meal Futures Market In China

The agricultural product futures market is an important place for regulating the marketprice and slowing down the price fluctuation of agricultural products. The fluctuation of itsprice is directly related to the prices on the spot market. It is particularly pronounced in thesoybean meal futures, because soybean meal has the characteristics of widely used and longindustrial chain. Thus, it has the vital practical significance of grasping the market featuresaccurately for maintaining stable prices of agricultural products, avoiding market risks andrealizing hedging. Chaos theory, as an important part of nonlinear science, can well reveal theordered structures and rules hidden behind the seemingly random economic phenomenon.Considering the complexity of futures market operation, the chaos theory provides a newangle for investigating the operation characteristics of soybean meal futures market.Based on nonlinear science, this paper will do the related research on the chaoscharacteristics of China soybean meal futures market from two aspects: one is chaosidentification and a second is chaos prediction.In the aspect of chaos identification, the paper selects fractal dimension and maximumLyapunov index on the basis of phase space reconstruction as the key indicators to identifythe chaos. The fractal dimension is calculated in non-integer form with G-P method and thepositive maximal Lyapunov exponent is obtained with small-data algorithm method, whichgives a qualitative proof for the chaotic characteristic of the soybean meal futures market inChina. Chaotic system is sensitive to its initial conditions and shows the local instability, so itis unpredictable for a long time. This is why the soybean meal futures market in China existmarket anomalies. However, the existence of chaotic attractors in chaos system ensures theglobal stability of the system, so the market fluctuations can achieve certain prediction for ashort term. Meanwhile, the paper analyses the cause of chaos in soybean meal futures marketin China from the three key points of chaos phenomenon, which is, nonlinearity, both orderand randomness, short-term predictable and long-term unpredictable co-exist.In the aspect of chaos prediction, on the one hand, we get251days the short-termprediction time cycle of soybean meal futures market through maximum Lyapunov index calculation, which is just a general estimate; on the other hand, considering the neuralnetwork can approximate nonlinear function completely, the neural network prediction modelwith chaos information is set up to fit and forecast the soybean meal futures price time series.The study shows that the soybean meal futures market is more suitable for short-termprediction compared with the long-term prediction and the accuracy will be greatly reducedafter250days. Moreover, the neural network prediction model shows a higher predictionprecision and it provides reference for building chaotic time series prediction model in thesense of methodology.In short, the paper demonstrated the chaotic characteristics of soybean meal futuresmarket in China from the chaos identification and prediction perspective. Based on the results,this paper discusses the reasons make futures market chaos and gets the followingenlightenment: For investors, they should take more attention to short-term economic impactto make the operation of hedging and risk aversion, when investing in futures products withchaos; For government regulators, they should build a healthy competitive investment climateto ensure the transparency of information, reduce market friction and improve the efficiencyof market operation; For scholars, based on the ideal predictive effect the neural networktechnology used in the study of time series with nonlinear characteristics get, they can makechaos identification of nonlinear time series to get their operation rules and predictable timescale grasp first, then set up the neural network prediction model with chaos information to fitand forecast the system. This method can greatly improve the precision and credibility of theprediction for the nonlinear system running tendency.