PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Probabilistic Temporal Process Model for Knowledge Processes: Handling a Stream of Linked Text
Marko Grobelnik, Dunja Mladenić and Jure Ferlez
In: SIKDD 2009, 12-16 Okt 2009, Ljubljana, Slovenia.

Abstract

The paper presents an approach to modelling the data obtained from an observed environment driven by knowledge processes. It is based on the proposed a formalism for presenting probabilistic temporal process model consisting of three major components: (1) background knowledge (in the form of ontologies), (2) observed data (in the form of a stream of data items represented in different data modalities and possibly enriched with background knowledge) and, (3) objectives to optimize (providing guidelines for analytic techniques). The goal is to enable maintaining a data structure - to store, summarize and respond to a wide variety of queries about the observed low level data and about information and knowledge derived from the process. The formalism is realized in software components. Its functioning is illustrated on three scenarios: personal email, corporate email and document collections. The resulting platform is called TNT (Text-Network-Time) according to the main data modalities being addressed within the software.

EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Information Retrieval & Textual Information Access
ID Code:6441
Deposited By:Jan Rupnik
Deposited On:08 March 2010