Semantic Analysis Based Approach for Relevant Text Extraction Using Ontology

Related Work

(Pal & Saha, 2014) have given an algorithm named simplified Lesk which is based on an unsupervised approach. This algorithm helps in extracting the summary from a large content of a document. The construction of summary is done for obtaining the important sentences with the help of conventional online dictionary. These significant sentences are assigned weights according to the importance of the content written by an author. Further these assigned weights are evaluated using the steps of the proposed simplified Lesk algorithm. This simplified Lesk algorithm only considered the keywords by using the WordNet which is not enough to extract the meaningful information from a document. So, the simplified Lesk algorithm is modified which considers the semantics extracted from conventional online dictionary like WordNet.

(Kulkarni & Apte, 2014) has given lexical chain computing algorithm in which the division of original documents is done into sentences using segmentation. Then a map is constructed for each sentence using four types of relations i.e. synonym, hypernym, hyponym, meronym. The distance of each word with another related word is calculated and a lexical chain is constructed which is assigned weight accordingly. The chain having highest weight is selected as the longest chain and the relevant text extraction is done depending on the relevant lexical chains obtained.

(Patil et al., 2004) used the statistical approach for extracting the relevant information from a document. The given approach can be used in application for retrieving the semantic association from a text depending on the query from the user in the form of keyword. In this paper, the process of extracting relevant information in the form of summary is done in various phases like extraction of data items, generation of query, documents gathering, creation of summary, and arrangement of summary to any user of a query.