Abstract
Extracting information from unstructured, brief and short text composed of short phrases, incomplete sentences, unordered sequence of words and words in short form not falling into any regular syntax is a challenging task. This paper describes an approach to automatically extract information from data rich unstructured text documents based on a domain dependent ontology and populate a database. Here, we apply pattern matching in terms of keywords/constants to extract the patterns and generate a structured text representation with respect to a domain specific ontology. The approach is illustrated on one such unstructured, short and brief text -classified matrimonial advertisement. The performance analysis of the approach on this case study is presented.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have