Micropinion generation

Abstract

This paper presents a new unsupervised approach to generating ultra-concise summaries of opinions. We formulate the problem of generating such a micropinion summary as an optimization problem, where we seek a set of concise and non-redundant phrases that are readable and represent key opinions in text. We measure representativeness based on a modified mutual information function and model readability with an n-gram language model. We propose some heuristic algorithms to efficiently solve this optimization problem. Evaluation results show that our unsupervised approach outperforms other state of the art summarization methods and the generated summaries are informative and readable.

Sentiment analysis: A review and comparative analysis of web services

“…On the other hand, in multi-document summarization once features and entities have been detected, the system has to group and/or order the different sentences which express sentiments related to those entities or features. The final summary can be presented as a graphic or a text showing the main features/entities and quantifying the sentiment with regard to each one in some way, for example, aggregating intensities of sentiments or counting the number of positive or negative sentences (Beineke et al, 2004;Pang and Lee, 2004;Park et al, 2012;Nishikawa et al, 2010b,a;Ganesan et al, 2010Ganesan et al, , 2012Tata and Di Eugenio, 2010). 4.…”

Discovering Related Clinical Concepts Using Large Amounts of Clinical Notes

“…Relatedness(SQ, RC i ) PROB is simply the probability that the search query and related concept occur together within a window of σ window . The second measure we used is a modified PMI measure, 5 which measures the strength of two concepts occurring together versus the two concepts occurring independently. The Relatedness(SQ, RC i ) score based on PMI is measured as follows: where X is computed based on Equation 1 as: …”

Three years - three moves - government verdicts to renovate customary 'Bharat' to contemporary 'India': evaluation of opinions from citizens

“…Key-phrase extraction uses a combination of statistical feature (based on weighted betweeness centrality scores of words) and co-location strength (based on nearest neighbour) (Kumar et al, 2016). The result of the analysis generates combination of words are principally a set of co-occurring word, and, while computing the ngram typically move one word forward (Ganesan et al, 2012). The generic representation of ngram is:…”