THEFT DETECTION VIA ADAPTIVE LEXICAL SIMILARITY ANALYSIS OF SOCIAL MEDIA DATA STREAMS
A method of detecting an event includes selecting a region and time frame of interest, obtaining a set of social media data streams associated with the region and the time frame of interest, and applying a lexical graph generation algorithm to the set of social media data streams to obtain lexical graphs. Performing similarity analysis on the lexical graphs is based on candidate lexical graphs related to the event to generate matching data, and investigating the event is based on the matching data.
1 . A computer-implemented method of detecting an event, the method comprising:
selecting a region and time frame of interest;
obtaining a set of social media data streams associated with the region and the time frame of interest;
applying, using a processor, a lexical graph generation algorithm to the set of social media data streams to obtain lexical graphs;
performing, using the processor, similarity analysis on the lexical graphs based on candidate lexical graphs related to the event to generate matching data; and
investigating the event based on the matching data.
2 . The computer-implemented method according to claim 1 , further comprising generating a candidate lexical graph dictionary that includes the candidate lexical graphs.
3 . The computer-implemented method according to claim 2 , wherein the generating the candidate lexical graph dictionary includes obtaining historical data that includes a known event.
4 . The computer-implemented method according to claim 3 , wherein the generating the candidate lexical graph dictionary further includes searching the historical data for a learning set of social media data streams.
5 . The computer-implemented method according to claim 4 , wherein the generating the candidate lexical graph dictionary further includes generating test lexical graphs from the learning set of social media data streams.
6 . The computer-implemented method according to claim 5 , wherein the generating the candidate lexical graph dictionary further includes selecting relevant ones of the test lexical graphs as the candidate lexical graphs.
7 . The computer-implemented method according to claim 1 , wherein the performing the similarity analysis includes matching words in the lexical graphs with words in the candidate lexical graphs.