A Word Sense Disambiguation Algorithm Based On Word Net Context

As the World Wide Web in the explosive growth of knowledge, vast amounts ofdata on the Web are stored in the form of natural language, so on a number ofknowledge acquisition, mining, information dissemination, and other naturallanguage processing NLP research fiery quickly, however, as stored in the knowledge" medium" on the existence of ambiguity, which makes the process of naturallanguage processing becomes complicated, difficult, and machine learning can notdetermine the ambiguity between languages. Thus, the word sense disambiguationwas born.WSD is a natural language processing a more important branch, it is also naturallanguage processing basic research, often contain: text processing, speechrecognition, machine translation MT, information access IE, information retrieval IR,dictionary lookup and data analysis and other fields. As the word sensedisambiguation is a natural language processing difficulties and focus, because wordsense disambiguation for other applications has important theoretical and practicalsignificance. Thus, word sense disambiguation task became middle, it has alsobecome an important process of natural language processing, the results of theirresearch directly applied to many aspects of information processing.WSD(Word Sense Disambiguation) work requires complete knowledgereasoning, and now due to the lack of access to knowledge has led to the result afterthe word sense disambiguation lead to reduced accuracy and coverage reduced, whichis also known as lack of access to knowledge leads bottlenecks occur, which not onlylimits the word sense disambiguation performance, reducing the scope of applicationof word sense disambiguation system. From another perspective, word sensedisambiguation can also determine the specific meaning of the word based on thecontext of words where, polysemy needed to determine must be the polysemy wherethe text in the context of being possible to achieve both a polysemy specific meaningis determined by its context, and the context in natural language computing wordprocessing, be sure to increase the time complexity and space complexity for thedirect impact when word sense disambiguation contexts where the word sense disambiguation results.WSD is also considered the AI-complete problem, this problem will have to betransformed to have knowledge of the document structure, again pre-defined rules todetermine the meaning of words based on the system to provide good knowledge base,based on a WordNet semantic context the key disambiguation WSD is to obtainknowledge in WordNet, the more knowledge if you get in the WordNet, the resultwill be more WSD ideal context of this paper, based on WordNet word sensedisambiguation algorithm is implemented as ambiguous words the establishment of asemantic graph, thereby providing a rich source of knowledge and credible wordsense disambiguation algorithm in this paper make up the bottleneck problem ofknowledge acquisition. In addition, the use of word sense disambiguation WordNetas the sole source of knowledge, reducing the problem to be marked meaningpolysemy, this makes sense disambiguation can be successfully applied to the contextof the search. Based on the above issues, this article will focus on how to obtain andbuild knowledge in semantic relations in WordNet map for polysemy as the maincontent of this paper, as WordNet development so far has been more than ten years ofhistory, which contains a wealth of knowledge. Method to determine the context ofthree ways: First, get the context of knowledge based on the context of the slidingwindow; Second, access to context-based dependency parsing tree-based knowledgeand access to the main research contents and results of context knowledge, this paper:one based on syntactic analysis tree to get contextual features of the word algorithm;analysis center to get the context of the characteristics of vocabulary words; performword sense disambiguation based on context. Second, the article uses the semanticrelationships between polysemy in WordNet, combined contexts word sensedisambiguation. Third herein algorithm to construct three kinds of semantic graphmodel, this semantic graph model WSD provides a large knowledge. Paper selectedexperimental set is Senseval-3tasks as word sense disambiguation, good results.