Identification Of The Associated Account Group In Futures Market

With the growing of Chinese futures market, the risk of the futures market is alsogrowing quickly. The risk of the futures market has predictability and measurability. Riskprediction means the estimatation and measurement of the risks. So they can determine thefrequency and intensity of the various risks. They can also provide basis with selecting theappropriate risk approach. Risk prediction generally includes the following two parts: one isto predict the probability of the risk, which means to find the regularity of the losses throughaccumulating and observating information; another one is to predict the strength of the risk,which means to lead to direct and indirect losses by supporsing the risk occurred.This paper describes the design and development of a sytem which establishes a set ofrisk control model by history data research which summarized by risk analysts in the dailywork. It can analyze the data based on history data. It can also analyze the relationshipbetween the data, so that it can control the risk in the future moe easily. In this case, it needsto develop a new system which can make the guests use the system more conveniently, andthe system should also analyze the history data quickly.The main contributions of this paper are as follows:(1) This paper designs the overall structure of the associated account group identificationsystem, which is divided into eight functional modules, namely, data conversion managementmodule, basic indicators data query scene management module, graphical managementmodule, Pearson correlation coefficients calculation management module, the associatedaccount group scene management module, the identification associated the account groupmanagement module, the secondary identification associated account group managemenetmodule and the associated account large group to re-identify management module, and therequirements, design and implementation of each of these functional modules are analyzed in detail;(2) By using one Data Extraction technology which named ETL(Extract-Transform-Load) and one multi-database middleware technology which namedDBROUTER, to extract clients’ position data, clients’ tradingday data and clients’ gains andlosses everyday. And these data will be transformed to be the basic clients’ data for thissystem. It provides the full data for Pearson correlation coefficients calculation and theidentification associated the account group.(3) Average value and standard deviation are calculated on the clients’ basic indicatorsdata. And it also calculates Pearson correlation coefficients with basic indicators data. Itspends less time after on optimizing the standard deviation algorithm and the Pearsoncorrelation coefficient algorithm.(4) Calculates the identification associated the account group for the related clients andDepth_First Search algorithm. And it spends less time on the identification associated theaccount group algorithm.(5) By changing conditions and changing related clients for the identified associatedaccount group, it can decrease conditions and exclude clients who are not need to be related.The identified associated account group will be identified for the second time, and it willincrease the results’ accurate.(6) Merging the secondary identification associated account groups and idenfying themagain, risk analysts can analyze the whole market risk.Finally, the operational status of the system is analyzed in detail from various angles, andthe system is successfully used in a real production environment.By using the identificationassociated the account group system, it can let risk analysts analysis history data more quicklyand more conveniently. And it can also protect the risk in the future. At the same time, thissystem design and implementation experience will give other project teams which need toanalyze other kinds of history data and real time data to play a very good reference.