International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 98 - Number 12 |
Year of Publication: 2014 |
Authors: Pratik Kamble, S. C. Dharamadhikari |
10.5120/17238-7572 |
Pratik Kamble, S. C. Dharamadhikari . Context Score based Term Weighting Model for Text Summarization. International Journal of Computer Applications. 98, 12 ( July 2014), 41-46. DOI=10.5120/17238-7572
Everybody is looking for relevant information briefly, which will cover information with small content. Summarization is the best for this. Current text summarization techniques do not consider the context i. e. background situation in that document. In this paper we are going to present the SentenceRank algorithm which will calculate the weight of the sentence based on the context score. We are going to make effective use of E-VSM : Enhance - Vector Space Model for bigram frequency count in whole corpus, where for each bigram we are going calculate the context score based on Bernoulli's model of randomness [1] [2]. Calculated bigrams context score is used in sentenceRank algorithm to calculate the context sensitive indexing weight of each sentence in a document. To reduce the redundancy in the sentences of summary, Cosine similarity measure is used to remove redundant sentence.