显示器电磁木马是通过控制计算机屏幕电磁辐射达到窃取信息目的的一种新型木马。当前的主流防护思想是用软件防护代替较为成熟但造价昂贵的硬件防护机制,然而目前软防护思想大多侧重于理论方法的探索,在实现机制上相对比较复杂。针对显...显示器电磁木马是通过控制计算机屏幕电磁辐射达到窃取信息目的的一种新型木马。当前的主流防护思想是用软件防护代替较为成熟但造价昂贵的硬件防护机制,然而目前软防护思想大多侧重于理论方法的探索,在实现机制上相对比较复杂。针对显示器电磁木马的工作特点提出了Soft-TEMPEST防护机制,设计了显示器电磁木马的ADFA(API Detection and Frequency Analysis)检测方法。该方法通过API函数序列的周期性挖掘分析,结合对屏幕像素信息的傅里叶变换及频谱分析,达到检测出木马进程的目的。测试结果表明,该方法能够成功检测出多种显示器电磁木马,而且原理简单,方便投入使用。展开更多
时空聚类算法是地理时空大数据挖掘的基础研究命题。针对传统CFSFDP聚类算法无法应用于时空数据挖掘的问题,本文提出一种时空约束的ST-CFSFDP(spatial-temporal clustering by fast search and find of density peaks)算法。在CFSFDP算...时空聚类算法是地理时空大数据挖掘的基础研究命题。针对传统CFSFDP聚类算法无法应用于时空数据挖掘的问题,本文提出一种时空约束的ST-CFSFDP(spatial-temporal clustering by fast search and find of density peaks)算法。在CFSFDP算法基础上加入时间约束,修改了样本属性值的计算策略,不仅解决了原算法单簇集多密度峰值问题,且可以区分并识别相同位置不同时间的簇集。本文利用模拟时空数据与真实的室内定位轨迹数据进行对比试验。结果表明,该算法在时间阈值90 s、距离阈值5 m的识别正确率高达82.4%,较经典ST-DBCSAN、ST-OPTICS及ST-AGNES聚类算法准确率分别提高了5.2%、4.2%和7.6%。展开更多
在OPTICS(ordering points to identify the clustering structure)算法主要考虑空间信息的基础上,提出了时空密度(STOPTICS)算法,增加了处理噪声孤立点时考虑时间距离的方法,并对每一兴趣区域内部的轨迹点根据时间轴做二度聚类,结合Apr...在OPTICS(ordering points to identify the clustering structure)算法主要考虑空间信息的基础上,提出了时空密度(STOPTICS)算法,增加了处理噪声孤立点时考虑时间距离的方法,并对每一兴趣区域内部的轨迹点根据时间轴做二度聚类,结合Apriori算法挖掘出用户频繁的行为模式,从而实现对用户兴趣区域及行为模式的挖掘研究。通过微软Geolife数据验证算法有效,为下一步处理用户轨迹数据奠定了基础。展开更多
Objective： To explore and analyze the points-selection rules in acupuncture treatment of mammary gland hyperplasia （MGH） by data mining and statistical method. Methods： Clinical literatures about the treatment of ...Objective： To explore and analyze the points-selection rules in acupuncture treatment of mammary gland hyperplasia （MGH） by data mining and statistical method. Methods： Clinical literatures about the treatment of MGH with acupuncture published in the recent 16 years were retrieved from Chinese Journal Full-text Database （CJFD） and established into a database by Excel. The SPSS 20 version software and Clementine 12.0 version software were adopted to analyze the frequency and association rules of points-selection in the treatment of MGH with acupuncture. Results： The top 3 points used most frequently in acupuncture treatment of MGH were Danzhon8 （CV 17）, Taichong （LR 3） and Zusanli （ST 36）; points from the Stomach Meridian of Foot Yangming and Liver Meridian of Foot Jueyin were most commonly used; the commonly selected points were predominantly distributed in thoracic and abdominal regions and lower limbs; emphasis on the combination use of local and distal points; of the specific points, the five Shu-Transmitting points were mostly used; association analysis showed that the associations among Taichong （LR 3）, Danzhong （CV 27） and Zusanli （ST 36） were the most significant. Conclusion： The data mining results substantially accord with the general rules of acupuncture-moxibustion theories in traditional Chinese medicine, able to reflect the points-selection principles and features in acupuncture treatment of MGH and provide evidence for the points selection in the treatment of MGH in acupuncture clinic.展开更多
文摘显示器电磁木马是通过控制计算机屏幕电磁辐射达到窃取信息目的的一种新型木马。当前的主流防护思想是用软件防护代替较为成熟但造价昂贵的硬件防护机制,然而目前软防护思想大多侧重于理论方法的探索,在实现机制上相对比较复杂。针对显示器电磁木马的工作特点提出了Soft-TEMPEST防护机制,设计了显示器电磁木马的ADFA(API Detection and Frequency Analysis)检测方法。该方法通过API函数序列的周期性挖掘分析,结合对屏幕像素信息的傅里叶变换及频谱分析,达到检测出木马进程的目的。测试结果表明,该方法能够成功检测出多种显示器电磁木马,而且原理简单,方便投入使用。
文摘时空聚类算法是地理时空大数据挖掘的基础研究命题。针对传统CFSFDP聚类算法无法应用于时空数据挖掘的问题,本文提出一种时空约束的ST-CFSFDP(spatial-temporal clustering by fast search and find of density peaks)算法。在CFSFDP算法基础上加入时间约束,修改了样本属性值的计算策略,不仅解决了原算法单簇集多密度峰值问题,且可以区分并识别相同位置不同时间的簇集。本文利用模拟时空数据与真实的室内定位轨迹数据进行对比试验。结果表明,该算法在时间阈值90 s、距离阈值5 m的识别正确率高达82.4%,较经典ST-DBCSAN、ST-OPTICS及ST-AGNES聚类算法准确率分别提高了5.2%、4.2%和7.6%。
文摘在OPTICS(ordering points to identify the clustering structure)算法主要考虑空间信息的基础上,提出了时空密度(STOPTICS)算法,增加了处理噪声孤立点时考虑时间距离的方法,并对每一兴趣区域内部的轨迹点根据时间轴做二度聚类,结合Apriori算法挖掘出用户频繁的行为模式,从而实现对用户兴趣区域及行为模式的挖掘研究。通过微软Geolife数据验证算法有效,为下一步处理用户轨迹数据奠定了基础。
基金supported by National Natural Science Foundation of China, No.81673979, No. 81473688, No. 81173265Supporting Program for New Century Excellent Talents in University of China, No.NCET-13-0827+4 种基金Administration of Traditional Chinese Medicine of Guangdong Province,No. 20141070Science and Technology Supporting Program of Guangzhou, No.2014J4100104Science and Technology Planning Project of Guangdong Province, No.2014A020212672Natural Science Foundation of Guangdong Province, No.2016A030313114, No. 2015A030313333Scientific Research and Innovation Fund of Jinan University/Fundamental Research Funds for the Central Universities, No. 21617467, No.21615464, No. 21615412~~
文摘Objective： To explore and analyze the points-selection rules in acupuncture treatment of mammary gland hyperplasia （MGH） by data mining and statistical method. Methods： Clinical literatures about the treatment of MGH with acupuncture published in the recent 16 years were retrieved from Chinese Journal Full-text Database （CJFD） and established into a database by Excel. The SPSS 20 version software and Clementine 12.0 version software were adopted to analyze the frequency and association rules of points-selection in the treatment of MGH with acupuncture. Results： The top 3 points used most frequently in acupuncture treatment of MGH were Danzhon8 （CV 17）, Taichong （LR 3） and Zusanli （ST 36）; points from the Stomach Meridian of Foot Yangming and Liver Meridian of Foot Jueyin were most commonly used; the commonly selected points were predominantly distributed in thoracic and abdominal regions and lower limbs; emphasis on the combination use of local and distal points; of the specific points, the five Shu-Transmitting points were mostly used; association analysis showed that the associations among Taichong （LR 3）, Danzhong （CV 27） and Zusanli （ST 36） were the most significant. Conclusion： The data mining results substantially accord with the general rules of acupuncture-moxibustion theories in traditional Chinese medicine, able to reflect the points-selection principles and features in acupuncture treatment of MGH and provide evidence for the points selection in the treatment of MGH in acupuncture clinic.
