Metadata based Text Mining for Generation of Side Information
Abstract
Text mining is knowledge analyzing technique to find a pattern. The side information is also called as metadata in most of the metadata based text mining applications. The side information consisting of large data in terms of weblogs, metadata, and non-textual data i.e. image/video, etc. This large data present in the unprocessed form which cannot be used for further text mining. Therefore, metadata based text mining algorithms are used to mine the useful information. In this paper, the proposed approach uses the different kind of pre-processing steps i.e. splitting, tokenize, steaming, parsing and chunking. For generating the side information i.e. title, name, affiliation, email address, place etc. a natural language processing (NLP) is used. To achieve the effective clustering, the proposed approach uses a classical partitioning method with a probabilistic model. The proposed approach is compared in terms of time required for mining of words, accuracy, and efficiency. The presented result shows that, the proposed approach performs better in terms of accuracy and running time. In future, a Security is provided for metadata based side information generation using Intrusion Detection System (IDS).
Related Papers
- Automated Metadata Extraction(2008)
- → A programmatic view of metadata, metadata services, and metadata flow in ATLAS(2012)1 cited
- → Statistical Metadata in Data Processing and Interchange(2005)2 cited
- Map Metadata - Going Beyond the Obvious/Connecting the Dots(2013)
- → Metadata(2011)