No icon

Linking Fine-Grained Locations in User Comments

Linking Fine-Grained Locations in User Comments


Many domain-specific websites host a profile page for each entity (e.g., locations on Foursquare, movies on IMDb, and products on Amazon) for users to post comments on. When commenting on an entity, users often mention other entities for reference or comparison. Compared with web pages and tweets, the problemof disambiguating the mentioned entities in user comments has not received much attention. This paper investigates linking fine-grained locations in Foursquare comments. We demonstrate that the focal location, i.e., the location that a comment is posted on, provides rich contexts for the linking task. To exploit such information, we represent the Foursquare data in a graph, which includes locations, comments, and their relations. A probabilistic model named FocalLink is proposed to estimate the probability that a user mentions a location when commenting on a focal location, by following different kinds of relations. Experimental results show that FocalLink is consistently superior under different collective linking settings.

Existing System:

For various types of texts like web pages and tweets, entity linking [1] has proved useful in facilitating understanding and searching those texts, as well as extracting information from them. In those texts, given a detected anchor that refers to some entity, an entity linker resolves the ambiguity of the anchor by mapping it to the right entry in some database.

We argue that investigations in location domain may give rise to techniques generalizable to the same linking problem in other domains. Although entity linking for formal documents like web pages has been relatively well studied, efforts on the same task for user comments remain limited.




Proposed System:

We address linking fine-grained locations in user comments, which may facilitate an effective use of user generated content and inspire similar tasks in other domains.

We propose a probabilistic linking model to exploit extra contextual information brought by focal locations.

We experimentally validate the superiority of our model under different collective linking settings.

The proposed solution deals with the short characteristics of user comments. To deal with the insufficiency of context for disambiguating mentions, we exploit the focal location and the relations between locations as extra contextual information. More importantly, the data graph enables estimating the possibility that a user mentions a location while commenting on another.

Comment As:

Comment (0)