Importance of Document Localization in Data m...

Importance of Document Localization in Data mining

Author Name : Sijin P, Champa H N

ABSTRACT

Document localization is the process of organizing documents on a database or repository and to retrieve them out by providing necessary information and proper workload based inference. Annotation is the process of adding meaningful notes or comments about an entity such as documents, attributes, data repositories and data spaces to make them more visible and expressive during document localization. In Machine learning oriented probabilistic data models it is used for generating values for attributes which could be used for generating a range of queries which are matched with a database schema. The proposed fuzzy document localization model (FDLM) lists out the top-k attributes by deriving a monotone fuzzy rank function based on query value Qval and content value Cval. The newly arrived documents are processed with the annotated documents which are conditionally modeled with ground truth attributes in a dynamic document categorization process. The semantic matches of attributes are identified by a pre-processed conceptualization framework this in turn increases the cardinality of result set. The system is biased with a biasing parameter β in order to maintain a balance with workload based query value and database oriented content value to set a selection bound over the range of accurate and approximate matches.

Keywords - Query value, Content Value, Confident segments, Annotation, Workload.