CS 730R: Topics in Data and Information Management
ثبت نشده
چکیده
1. Summary. The paper proposes a novel streaming algorithm to mine the top-k episodes in a stream of events. The frequency of the episodes is computed over a sliding window which length is defined by the user. The key idea in this paper is based on two new concepts related to the stream: maximum rate of change and top-k separation. The sliding window is decomposed into batches and the previous two concepts are used in tracking the top-k elements in every batch in the stream. The author show how an exact and approximate solution to the top-k problem can be found in the stream by mining only a fraction of the frequent episodes in each batch. This approach is the first solution for mining events in a stream and represents a generalization of the existing streaming pattern discovery algorithms.
منابع مشابه
CS 730R: Topics in Data and Information Management
1. Summary. The paper presents a pattern mining approach to mine recent temporal patterns in multivariate time series. The major contribution consists in learning events from time series which is done via mapping time series into state sequences and mining from the transformed sequence the recent patterns to use for SVM. The authors show how their framework allows to efficiently perform mining ...
متن کاملCS 730R: Topics in Data and Information Management
1. Summary. In this paper the authors propose a differentially privacy preserving algorithm for mining frequent itemset. This work differs from the other privacy preserving miners present in literature, indeed this algorithm mines the itemset by enforcing cardinality constraints on the transactions present in the dataset. In particular the authors study how the reduction the cardinality of the ...
متن کاملCS 730R: Topics in Data and Information Management
1. Summary. The paper is focused on the problem of answering a set of queries on sliding windows over a data stream. The setting is very appealing and suitable for many applications where the queries are windowbased rather than on the entire stream. This work introduces the concept of window and query workload that were missing in the state of the art of differentially private streaming algorit...
متن کاملCS 730R: Topics in Data and Information Management
1. Summary. In this paper the authors presents two major contributions to the recent work of privacy on streaming data initially introduced by Dwork et al. [2]. First the authors study the family of problems related to the decayed sums. These problems are motivated by the fact that only the most recent part of the stream is often considered important in real applications rather than considering...
متن کاملResearch Priorities of Iranian Faculty Members and PhD Students in Medical Library and Information Science
Objective: The field of “medical library and information science” has a broad research platform due to its interdisciplinary nature. However, the lack of resources necessitates the attention to research priorities in this field. This study aims to identify the research priorities of Iranian faculty members and PhD students in medical library and information science. Methods: This is a descript...
متن کامل