期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2015
卷号：4
期号：3
页码：750-753
出版社：Shri Pannalal Research Institute of Technolgy
摘要：There are numerous techniques available to find occurrences of activities in time-stamped observation data with each occurrences having an associated probability.For the entire research community the activity recognition is considered to be a big challenge. The existing techniques cannot deal with "zero day" attacks that have never seen before .In the proposed system it is able to find the subsequenes of the observation data,called unexplained sequence,that known models are not able to "explain" with a certain confidence.Thus we have to consider a known set of activities which contains both innocuous and dangerous activities .We want to monitor and also we want to identify the unexplained subsequences in an observation sequence that are poorly explained that is,we want to identify the activities that are not present in this predefined set.Top-k algorithms are used to identify the top-k totally and partially unexplained activities.In the proposed system these algorithms are applied on the Cyber Security data sets. These algorithms are more efficient and provides faster search for identifying totally and partially unexplained sequences.
