现实量化交易应用中,传统的模糊数据挖掘算法往往需要针对给定的量化交易设定最小支持度阈值,然而,这些方法中存在的普遍问题是很难找到合适的最小支持度阈值,并且因为推导出的规则通常是常识而没有实际的商业意义。为了解决这个问题,...现实量化交易应用中,传统的模糊数据挖掘算法往往需要针对给定的量化交易设定最小支持度阈值,然而,这些方法中存在的普遍问题是很难找到合适的最小支持度阈值,并且因为推导出的规则通常是常识而没有实际的商业意义。为了解决这个问题,提出了一种无需最小支持度阈值的模糊关联规则(fuzzy coherent rule,FCR)挖掘算法。首先将量化交易转换成模糊集,然后通过收集已经生成的模糊集生成候选模糊关联规则,最后计算出列联表并用其检查这些候选模糊关联规则是否满足四项判断准则。如果满足,则可以确定为模糊关联规则。在Foodmart数据集上的实验验证了所提算法的有效性,相比原始模糊关联规则(fuzzy association rules,FAR)挖掘算法,所提的FCR方法能够推导出更多的规则,并且能够在高置信度时推导出更多有用的规则。展开更多
A precise localization for mobile target in wireless sensor networks is presented in this letter,where a geometrical relationship is explored to improve the location estimation for mobile target,in-stead of a simple c...A precise localization for mobile target in wireless sensor networks is presented in this letter,where a geometrical relationship is explored to improve the location estimation for mobile target,in-stead of a simple centroid approach.The equations of location compensation algorithm for mobiletarget are derived based on linear trajectory prediction and sensor selective activation.The resultsbased on extensive simulation experiments show that the compensation algorithm gets better per-formance in metrics of quality of tracking and energy efficiency with the change of sensor sensing range,the ratio of sensing range and sensor activation range,and the data sampling rate than traditionalmethods,which means our proposing can achieve better quality-energy tradeoff for mobile target inwireless sensor networks.展开更多
This paper proposes a double Markov model of the double continuous auction for describing intra-day price changes. The model splits intra-day price changes as the repetition of one tick price moves and assumes order a...This paper proposes a double Markov model of the double continuous auction for describing intra-day price changes. The model splits intra-day price changes as the repetition of one tick price moves and assumes order arrivals are independent Poisson random processes. The dynamic process of price formation is described by a birth-death process of the double M/M/1 server queue corresponding to the best bid/ask. The initial depths of the best bid and ask are defined as different constants depending on the last price change. Thus, the price changes in the model follow a first-order Markov process. As the initial depth of the best bid/ask is originally larger than that of the opposite side when the last price is down/up, the model may explain the negative autocorrelations of the price of the best bid/ask. The estimated parameters are based on the real tick-by-tick data of the Nikkei 225 futures listed in Osaka Stock Exchanges. The authors find the model accurately predicts the returns of Osaka Stock Exchange average.展开更多
文摘现实量化交易应用中,传统的模糊数据挖掘算法往往需要针对给定的量化交易设定最小支持度阈值,然而,这些方法中存在的普遍问题是很难找到合适的最小支持度阈值,并且因为推导出的规则通常是常识而没有实际的商业意义。为了解决这个问题,提出了一种无需最小支持度阈值的模糊关联规则(fuzzy coherent rule,FCR)挖掘算法。首先将量化交易转换成模糊集,然后通过收集已经生成的模糊集生成候选模糊关联规则,最后计算出列联表并用其检查这些候选模糊关联规则是否满足四项判断准则。如果满足,则可以确定为模糊关联规则。在Foodmart数据集上的实验验证了所提算法的有效性,相比原始模糊关联规则(fuzzy association rules,FAR)挖掘算法,所提的FCR方法能够推导出更多的规则,并且能够在高置信度时推导出更多有用的规则。
基金the Joint Funds of Guangdong-NSFC(U0735003)the National Natural Science Foundation of China(No.60604029,60702081)+2 种基金the Natural ScienceFoundation of Zhejiang Province(No.Y106384)the Sci-ence and Technology Project of Zhejiang Province(No.2007C31038)and the Scientific Research Fund of Zhejiang Provincial Education(No.20061345).
文摘A precise localization for mobile target in wireless sensor networks is presented in this letter,where a geometrical relationship is explored to improve the location estimation for mobile target,in-stead of a simple centroid approach.The equations of location compensation algorithm for mobiletarget are derived based on linear trajectory prediction and sensor selective activation.The resultsbased on extensive simulation experiments show that the compensation algorithm gets better per-formance in metrics of quality of tracking and energy efficiency with the change of sensor sensing range,the ratio of sensing range and sensor activation range,and the data sampling rate than traditionalmethods,which means our proposing can achieve better quality-energy tradeoff for mobile target inwireless sensor networks.
基金supported by the National Natural Science Foundation of China under Grant Nos.71173060,71031003the Fundamental Research Funds for the Central Universities under Grant No.HIT.HSS.201120partially supported by JSPS KAKENHI under Grant No.22560059
文摘This paper proposes a double Markov model of the double continuous auction for describing intra-day price changes. The model splits intra-day price changes as the repetition of one tick price moves and assumes order arrivals are independent Poisson random processes. The dynamic process of price formation is described by a birth-death process of the double M/M/1 server queue corresponding to the best bid/ask. The initial depths of the best bid and ask are defined as different constants depending on the last price change. Thus, the price changes in the model follow a first-order Markov process. As the initial depth of the best bid/ask is originally larger than that of the opposite side when the last price is down/up, the model may explain the negative autocorrelations of the price of the best bid/ask. The estimated parameters are based on the real tick-by-tick data of the Nikkei 225 futures listed in Osaka Stock Exchanges. The authors find the model accurately predicts the returns of Osaka Stock Exchange average.
