A hidden Markov model information retrieval system
Proceedings of the 22nd annual international ACM SIGIR conference on Research and development in information retrieval
SemTag and seeker: bootstrapping the semantic web via automated semantic annotation
WWW '03 Proceedings of the 12th international conference on World Wide Web
A Semantic Web Primer
GATE: an architecture for development of robust HLT applications
ACL '02 Proceedings of the 40th Annual Meeting on Association for Computational Linguistics
| Hi-index | 0.00 |
The vision of Semantic Web can be realized when there are masses of machine-processable semantic metadata. Manual construction of metadata is not feasible, methods for automated semantic annotation have been developed. Semantic annotation is the process to identify the most appropriate semantic tagging to entities. The key challenge in automatic semantic annotation is resolving ambiguities in identifying semantic tagging. We propose the OntoNEO (Ontology-based semantic Named Entity disambiguatiOn), which is a new named entity disambiguation algorithm, based on Hidden Markov Model (HMM) and semantic contexts. OntoNEO employs HMM to represent probabilistic model of a named entities among various sentences. For each named entity in ontology, we build a probabilistic model using HMM from corpus of documents. The sequence of named entities are generated using HMM. The probabilistic models of the named entity have been created considering the sequence, and then it used to resolve ambiguity of a named entity for semantic tagging from the perspective of sequential probability. We generate appropriate semantic contexts from a text to understand the meaning of a semantically ambiguous named entity and solve the problem of ambiguities during semantic annotation by searching the optimized Markov Model corresponding to the sequence of words included in the semantic contexts. Based on our OntoNEO, ubiquitous applications can reason about named entity free of contradictions. Compared to SemTag algorithm, our system has an improved performance by about 18%.