Active-passive Modes And Space Granularity Optimization In The Regulation Of Bulk Commodity Trading Market

Bulk commodity trading market involves national strategic materials such as energy,minerals,cotton,grain and oil,which is characterized by large transaction volume,large price fluctuation,large transaction risk,and large impact radiation.It is related to the lifeblood of national economic development,and has always been the key concern of the regulatory authorities.The vast majority of cases in commodities trading market show that many risk events are caused by the trading behavior of abnormal trading subject,therefore,how to effectively supervise abnormal trading subjects in the commodity trading market in order to maintain the stability and development of the commodity trading markets has become the main issues facing the regulation of the commodity trading market at present.Regulatory model issues and regulatory space granularity issues are currently two types of regulatory issues in the bulk commodity trading market.In the bulk commodity trading market,the regulatory mode can be divided into passive regulatory mode and active regulatory mode.The passive regulatory mode usually carries out supervision after the occurrence of market risk events,which has the advantages of high regulatory efficiency and less regulatory tasks,but has serious regulatory hysteresis.The active supervision mode can carry out supervision in advance,but the current active supervision mode relies heavily on manual supervision,which has the problems of heavy supervision task burden and low supervision efficiency.In the granularity of commodity regulatory space,the current single regulatory scope cannot effectively target cross-platform transactions,leading to the fact that trading subjects can avoid market supervision by dispersing their trading behaviors.In order to overcome the problems of regulatory lag and heavy regulatory burden in a single regulatory mode and the inability to effectively supervise cross-platform trading behavior in a single spatial granularity,this thesis conducts an in-depth study on the optimization of active and passive regulatory modes and regulatory spatial granularity in bulk commodity trading markets.In the research of active and passive supervision mode,this thesis proposes an anomaly detection algorithm based on power law distribution to solve the problems of low supervision efficiency and heavy supervision task in current active supervision mode and hysteresis in passive supervision mode.The algorithm constructs the trading data into the trading network model,and can separate the abnormal trading subjects from the trading data according to the behavior characteristics of the abnormal trading in the bulk commodity trading market.The finally experimental results show that the anomaly detection algorithm proposed in this thesis can effectively mine the abnormal trading subjects in the unlabeled bulk commodity trading data.At the same time,it has a higher abnormal transaction subject accuracy as well as full detection rate compared with other unsupervised algorithms.This shows in commodities regulation mode using the algorithm to report the transaction data of anomaly detection,and then the supervisory department supervises the abnormal transaction subjects,which can ensure that the supervisory department can supervise the market in advance before the outbreak of risk events,this solves the problem of regulatory lag in the passive supervision model.It can also greatly reduce the number of transaction subjects that the regulatory authorities need to analyze in active supervision,so as to effectively solve the problems of heavy supervision burden and low supervision transaction existing in the current active supervision mode.In the research of spatial granularity,this thesis proposes a spatial granularity recommendation algorithm based on community discovery to solve the problem that the crossplatform behavior of transaction users cannot be effectively monitored under the single regulatory spatial granularity.The algorithm first trading platform and trading in commodities trading market user modeling into binary heterogeneous network,and then projects the bipartite network based on the cross-platform behavior of common users,and finally uses modularity as the optimization objective to divide the community structure of the projection network,obtaining the granularity is recommended in the regulatory space of each trading plat form.The simulation results show that,under the background of spatial granularity research of commodity regulation,the spatial granularity results obtained by the community discovery algorithm based on projection method adopted in this thesis are better than the comparison algorithm in terms of indicators such as network modularity,standard mutual information and running time of trading platform.This suggests that in the commodities space size optimization using the proposed algorithm can better divide the trading platform of cross-platform trading under the same regulatory space granularity,granularity can't solve the single regulation space regulation cross-platform trading behavior problems,on the basis of avoiding the global regulatory,regulators can effectively regulated cross-platform transaction behavior.In order to improve the regulatory efficiency and quality of regulatory departments.At last,based on the research of the first two points,the design and development of the prototype system of active and passive regulation and spatial granularity optimization in the bulk commodity trading market are completed.In the prototype system design and development,this thesis first proposed the overall design of active and passive supervision and spatial granularity prototype system,and then describes the detailed design of each module according to the overall design.Finally,the system is tested and analyzed by black-box testing in terms of both system function and performance to prove the effectiveness and stability of the system operation.