Coreference Resolution to Support IE from Indian Classical Music Forums

Proceedings of Recent Advances in Natural Language Processing, pages 91–96, Hissar, Bulgaria, Sep 7–9 2015. Coreference Resolution to Support IE from Indian Classical Music Forums Joe Cheri Ross Pushpak Bhattacharyya Department of Computer Science and Engineering Indian Institute of Technology Bombay {joe,pb}@cse.iitb.ac.in Abstract Efficient music information retrieval (MIR) require to have meta information about music along with content based information in the knowledge base. Dis- cussion forums on music are rich sources of information gathered from a wider audience. Taking into consideration the nature of text in these web resources, the yield of relation extraction is quite depen- dent on resolving the entity references in the document. Among the few music fo- rums dealing with Indian classical music, rasikas.org (rasikas, 2015) having rich information about artistes, raga and other music concepts is taken for our study. The forum posts generally contain anaphoric references to the main topic of the thread or any other entity in the discourse. In this paper we focus on coreference resolution for short discourse noisy text like that of forum posts. Since grammatical roles capture relation between mentions in a discourse, those features extracted from dependency parsing are widely explored along with semantic compatibility feature. On investigation of issues, the need for integrating known dependencies between features emerged. A Bayesian network with predefined network structure evaluated, since a Bayesian belief networkis enacts a probabilistic rule based system. To the extent possible the superior be- haviour of Bayesian network over SVM is analysed. 1 Introduction Information extraction from music repositories in- volves analysis of music audio. Efficient extrac- tion of music information require meta-data along with content based information. The need for metadata led to information extraction from blogs and forums related to music. This should con- tain information about artistes, performances, mu- sic concepts etc. Apart from the available litera- ture about Indian classical music, there are a few forums and blogs having rich metadata. Extract- ing information from these sources help to aug- ment music ontology for Indian classical music with meta information along with content based information. Among the two main divisions in In- dian classical music, Carnatic music community is more involved in web based discussions and infor- mation dissemination. Rasikas.org (rasikas, 2015) is one among the prominent discussion forums where they have discussions pertaining to Carnatic music topics comprising ragas, talas, artistes etc. Extracting information from unstructured noisy text in websites of this kind is quite challenging. Efficient extraction of relations also require reso- lution of entities in the documents. Apart from re- solving the entities with the real world entities, the intra-relations between the entities within the dis- course have to be resolved. Identification of enti- ties is a critical step in information extraction fol- lowed by identification of relations between them. Posts in most forums are written in informal language with pronominal and alias mentions re- ferring to the main topic of discussion or to an- other related entity mentioned in the discourse. Effecient extraction of relation is dependent on …

Content truncated to 500 words