We study the combination of symbol frequence analysis and negative selection for anomaly detection of discrete sequences where conventional negative selection algorithms are not practical due to data sparsity.Theoreti...We study the combination of symbol frequence analysis and negative selection for anomaly detection of discrete sequences where conventional negative selection algorithms are not practical due to data sparsity.Theoretical analysis on ergodic Markov chains is used to outline the properties of the presented anomaly detection algorithm and to predict the probability of successful detection.Simulations are used to evaluate the detection sensitivity and the resolution of the analysis on both generated artificial data and real-world language data including the English Wikipedia.Simulation results on large reference corpora are used to study the effects of the assumptions made in the theoretical model in comparison to real-world data.展开更多
The time-frequency analysis and anomaly detection of wavelet transformation make the method irresistibly advan- tageous in non-stable signal processing. In the paper, the two characteristics are analyzed and demonstr...The time-frequency analysis and anomaly detection of wavelet transformation make the method irresistibly advan- tageous in non-stable signal processing. In the paper, the two characteristics are analyzed and demonstrated with synthetic signal. By applying wavelet transformation to deformation data processing, we find that about 4 months before strong earthquakes, several deformation stations near the epicenter received at the same time the abnormal signal with the same frequency and the period from several days to more than ten days. The GPS observation sta- tions near the epicenter all received the abnormal signal whose period is from 3 months to half a year. These ab- normal signals are possibly earthquake precursors.展开更多
传统的窃电检测方法大多直接在原始功率序列上构建模型,且没有同时考虑长时序列中的周期依赖关系以及周期之间的局部关联特征,限制了电力用户行为规律的深层挖掘。提出了一种综合采用时频模态融合和多尺度特征提取的高精度窃电检测模型...传统的窃电检测方法大多直接在原始功率序列上构建模型,且没有同时考虑长时序列中的周期依赖关系以及周期之间的局部关联特征,限制了电力用户行为规律的深层挖掘。提出了一种综合采用时频模态融合和多尺度特征提取的高精度窃电检测模型。采用经验模态分解方法,将原始信号分解为多个本征模态信号和一个残差信号,依据模糊熵值与皮尔逊相关系数找到同时包含局部信息与原始信号信息较多的模态,并将选择的模态信号与原信号进行拼接,这样既可以提升模型的维度,又能放大窃电用户与正常用户的局部差异;将拼接好的数据先输入卷积神经网络进行局部特征提取,并从提取到的特征输入多头自注意力机制神经网络模型中提取全局特征,从而实现多尺度特征提取,以增强模型提取特征的适应性。在公开数据集上的实验结果表明,所提模型的F1值达到了76.71%、召回率达到了87.99%、曲线下面积(area under the curve,AUC)值达到了93.11%,相比于现有方法均取得了明显提升。展开更多
Android is the mobile operating system most frequently targeted by malware in the smartphone ecosystem,with a market share significantly higher than its competitors and a much larger total number of applications.Detec...Android is the mobile operating system most frequently targeted by malware in the smartphone ecosystem,with a market share significantly higher than its competitors and a much larger total number of applications.Detection of malware before being published on official or unofficial application markets is critically important due to the typical end users'widespread security inadequacy.In this paper,a novel feature selection method is proposed along with an Android malware detection approach.The feature selection method proposed in this study makes use of permissions,API calls,and strings as features,which are statically extractable from the Android executables(APK files)and it can be used in a machine learning process with different algorithms to detect malware on the Android platform.A novel document frequency-based approach,namely Delta IDF,was designed and implemented for feature selection.Delta IDF was tested upon three universal benchmark datasets that contain Android malware samples and highly promising results were obtained by using several binary classification algorithms.展开更多
基金funded by the Academy of Finland under Grant No.214144
文摘We study the combination of symbol frequence analysis and negative selection for anomaly detection of discrete sequences where conventional negative selection algorithms are not practical due to data sparsity.Theoretical analysis on ergodic Markov chains is used to outline the properties of the presented anomaly detection algorithm and to predict the probability of successful detection.Simulations are used to evaluate the detection sensitivity and the resolution of the analysis on both generated artificial data and real-world language data including the English Wikipedia.Simulation results on large reference corpora are used to study the effects of the assumptions made in the theoretical model in comparison to real-world data.
基金Joint Seismological Science Foundation of China (604021) and National Natural Science Foundation of China(40074024).
文摘The time-frequency analysis and anomaly detection of wavelet transformation make the method irresistibly advan- tageous in non-stable signal processing. In the paper, the two characteristics are analyzed and demonstrated with synthetic signal. By applying wavelet transformation to deformation data processing, we find that about 4 months before strong earthquakes, several deformation stations near the epicenter received at the same time the abnormal signal with the same frequency and the period from several days to more than ten days. The GPS observation sta- tions near the epicenter all received the abnormal signal whose period is from 3 months to half a year. These ab- normal signals are possibly earthquake precursors.
文摘传统的窃电检测方法大多直接在原始功率序列上构建模型,且没有同时考虑长时序列中的周期依赖关系以及周期之间的局部关联特征,限制了电力用户行为规律的深层挖掘。提出了一种综合采用时频模态融合和多尺度特征提取的高精度窃电检测模型。采用经验模态分解方法,将原始信号分解为多个本征模态信号和一个残差信号,依据模糊熵值与皮尔逊相关系数找到同时包含局部信息与原始信号信息较多的模态,并将选择的模态信号与原信号进行拼接,这样既可以提升模型的维度,又能放大窃电用户与正常用户的局部差异;将拼接好的数据先输入卷积神经网络进行局部特征提取,并从提取到的特征输入多头自注意力机制神经网络模型中提取全局特征,从而实现多尺度特征提取,以增强模型提取特征的适应性。在公开数据集上的实验结果表明,所提模型的F1值达到了76.71%、召回率达到了87.99%、曲线下面积(area under the curve,AUC)值达到了93.11%,相比于现有方法均取得了明显提升。
文摘Android is the mobile operating system most frequently targeted by malware in the smartphone ecosystem,with a market share significantly higher than its competitors and a much larger total number of applications.Detection of malware before being published on official or unofficial application markets is critically important due to the typical end users'widespread security inadequacy.In this paper,a novel feature selection method is proposed along with an Android malware detection approach.The feature selection method proposed in this study makes use of permissions,API calls,and strings as features,which are statically extractable from the Android executables(APK files)and it can be used in a machine learning process with different algorithms to detect malware on the Android platform.A novel document frequency-based approach,namely Delta IDF,was designed and implemented for feature selection.Delta IDF was tested upon three universal benchmark datasets that contain Android malware samples and highly promising results were obtained by using several binary classification algorithms.
