관심 지점 탐지를 위한 모호성 해소 방법

다국어 초록 (Multilingual Abstract)

A point of interest (POI) is a term referring to a geographic entity such as a landmark, transportation, restaurant, annual festival and even a region. POIs are the basis for management and linking of the data in location-based applications. In this paper we propose a two-step method for POI detection from text. The first step is POI candidate detection based on the approximate entity recognition method that does not need any training data. The second step is POI candidate disambiguation that selects the best ontological entity for the POI candidate. We propose two POI candidate disambiguation methods: one based on theme locations and the other based on POI graphs. Furthermore, we construct a test collection for effectiveness evaluation and analyze the advantages and disadvantages of each POI candidate disambiguation method.

참고문헌 (Reference)

1 이향진, "모바일 사용자의 대용량 GPS 데이터로부터 시맨틱 관심지점 발견" 한국정보과학회 39 (39): 175-184, 2012

2 최정화, "대용량의 GPS 데이터 분석을 위한 사용자 선호 경로 모델 표현" 한국정보과학회 39 (39): 315-327, 2012

3 N. O’Hare, "Modeling locations with social media" 2012

4 A. Rae, "Mining the Web for Points of Interest" 2012

5 Li, G., "Faerie: efficient filtering algorithms for approximate dictionary-based entity extraction" 529-540, 2011

6 Wang, W., "Efficient approximate entity extraction with edit distance constraints" 759-770, 2009

7 Wang, "Detecting Geographic Locations from Web Resources" 2005

8 Ding, J., "Computing geographic scopes of web resource" 2000

9 Han, X., "Collective Entity Linking in Web Text: A Graph-Based Method" 2011

10 Deng, D., "An Efficient Trie-based Method for Approximate Entity Extraction with Edit-Distance Constraints" 762-773, 2012