The main purpose of this paper is to generalize the study of the Hecke-Rogers type series,which are the extensions of truncated theorems obtained by Andrews,Merca,Wang and Yee.Our proofs rely heavily on the theory of ...The main purpose of this paper is to generalize the study of the Hecke-Rogers type series,which are the extensions of truncated theorems obtained by Andrews,Merca,Wang and Yee.Our proofs rely heavily on the theory of Bailey pairs.展开更多
基于差分隐私的时间序列模式挖掘方法中,序列的最大长度以及添加拉普拉斯噪声的多少直接制约着挖掘结果的可用性.针对现有时间序列模式挖掘方法全局敏感度过高、挖掘结果可用性较低的不足问题,提出了一种基于序列格的差分隐私下时间序...基于差分隐私的时间序列模式挖掘方法中,序列的最大长度以及添加拉普拉斯噪声的多少直接制约着挖掘结果的可用性.针对现有时间序列模式挖掘方法全局敏感度过高、挖掘结果可用性较低的不足问题,提出了一种基于序列格的差分隐私下时间序列模式挖掘方法PrivTSM(Differentially Private Time Series Pattern Mining).该方法首先利用最长路径的策略对原始数据库进行截断处理;在此基础上,采用表连接操作生成满足差分隐私的序列格;结合序列格结构本身的特性,合理分配隐私预算,提高输出模式的可用性.理论分析表明PrivTSM方法满足ε-差分隐私,基于真实数据库上实验结果表明,PrivTSM方法的准确率TPR(True Postive Rate)和平均相对误差ARE(Average Relative Error)明显优于N-gram和Prefix-Hybrid方法.展开更多
The analytical renewal function(RF) is not tractable of the exponential Weibull(EW) distribution. In the proposed model, the n-fold convolution of the EW cumulative distribution function(CDF) is approximated by a n-fo...The analytical renewal function(RF) is not tractable of the exponential Weibull(EW) distribution. In the proposed model, the n-fold convolution of the EW cumulative distribution function(CDF) is approximated by a n-fold convolutions of Gamma and normal CDFs. We obtain the EW RF by a series approximation model. The method is very simple in the computation. When the parameters are unknown, we present the asymptotic confidence interval of the RF. The validity of the asymptotic confidence interval is checked via some numerical experiments.展开更多
基金Supported by the National Natural Science Foundation of China(11871370 and 12001182)the Fundamental Research Funds for the Central Universities(531118010411)。
文摘The main purpose of this paper is to generalize the study of the Hecke-Rogers type series,which are the extensions of truncated theorems obtained by Andrews,Merca,Wang and Yee.Our proofs rely heavily on the theory of Bailey pairs.
文摘基于差分隐私的时间序列模式挖掘方法中,序列的最大长度以及添加拉普拉斯噪声的多少直接制约着挖掘结果的可用性.针对现有时间序列模式挖掘方法全局敏感度过高、挖掘结果可用性较低的不足问题,提出了一种基于序列格的差分隐私下时间序列模式挖掘方法PrivTSM(Differentially Private Time Series Pattern Mining).该方法首先利用最长路径的策略对原始数据库进行截断处理;在此基础上,采用表连接操作生成满足差分隐私的序列格;结合序列格结构本身的特性,合理分配隐私预算,提高输出模式的可用性.理论分析表明PrivTSM方法满足ε-差分隐私,基于真实数据库上实验结果表明,PrivTSM方法的准确率TPR(True Postive Rate)和平均相对误差ARE(Average Relative Error)明显优于N-gram和Prefix-Hybrid方法.
基金supported by the National Natural Science Foundation of China(71801186)the National Natural Science Foundation of Guangdong(2018A030313829)+2 种基金the Science and Technology Innovation Guidance Project of Zhaoqing,Guangdong Province(201804031503)the higher education colleges and universities innovation strong school project of Guangdong(2016KTSCX153)the teaching reform project of Zhaoqing University(zlgc201745)
文摘The analytical renewal function(RF) is not tractable of the exponential Weibull(EW) distribution. In the proposed model, the n-fold convolution of the EW cumulative distribution function(CDF) is approximated by a n-fold convolutions of Gamma and normal CDFs. We obtain the EW RF by a series approximation model. The method is very simple in the computation. When the parameters are unknown, we present the asymptotic confidence interval of the RF. The validity of the asymptotic confidence interval is checked via some numerical experiments.
