Extracting significant time varying features from text
Proceedings of the eighth international conference on Information and knowledge management
Automatic generation of overview timelines
SIGIR '00 Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval
Efficient Similarity Search In Sequence Databases
FODO '93 Proceedings of the 4th International Conference on Foundations of Data Organization and Algorithms
An Online Algorithm for Segmenting Time Series
ICDM '01 Proceedings of the 2001 IEEE International Conference on Data Mining
PKDD '97 Proceedings of the First European Symposium on Principles of Data Mining and Knowledge Discovery
Time Series Segmentation for Context Recognition in Mobile Devices
ICDM '01 Proceedings of the 2001 IEEE International Conference on Data Mining
Indexing multi-dimensional time-series with support for multiple distance measures
Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining
On lossy time decompositions of time stamped documents
Proceedings of the thirteenth ACM international conference on Information and knowledge management
Information Preserving Time Decompositions of Time Stamped Documents*
Data Mining and Knowledge Discovery
Efficient algorithms for segmentation of item-set time series
Data Mining and Knowledge Discovery
An approach for temporal analysis of email data based on segmentation
Data & Knowledge Engineering
Extracting hot spots of topics from time-stamped documents
Data & Knowledge Engineering
Hi-index | 0.00 |
Identifying temporal information of topics from a document set typically involves constructing a time decomposition of the time period associated with the document set. In an earlier work, we formulated several metrics on a time decomposition, such as size, information loss, and variability, and gave dynamic programming based algorithms to construct time decompositions that are optimal with respect to these metrics. Computing information loss values for all subintervals of the time period is central to the computation of optimal time decompositions. This paper proposes several algorithms to assist in more efficiently constructing an optimal time decomposition. More efficient, parallelizable algorithms for computing loss values are described. An efficient top-down greedy heuristic to construct an optimal time decomposition is also presented. Experiments to study the performance of this greedy heuristic were conducted. Although lossy time decompositions constructed by the greedy heuristic are suboptimal, they seem to be better than the widely used uniform length decompositions.