Multi-document summarization

About: Multi-document summarization is a research topic. Over the lifetime, 2270 publications have been published within this topic receiving 71850 citations.

Research on Multi-document Summarization Using Lexical Cohesion

Abstract: This paper investigates using lexical cohesion to generate a moderately fluent semantic summary from a collection of documents written in Chinese. Based on the algorithm of cohesion analysis using the relationship among the words in the HowNet knowledge database, the built system computes concept frequency rather than word frequency as a measurement of importance. It merges the analysis of lexical semantics and some summarization principles to remove the redundancy and remain the difference in multiple documents. Such approach reduces information loss due to vocabulary switching in the summarization process and the use of a more general notion of relatedness which is based on lexical semantics. Thus we can take into account some more-distant relationship between words. Evaluation results show that the performance of the presented system is obviously better than that of the baseline system. The system can be applied to on-line web texts processing.

Study on Dynamic Multi-document Summarization System Framework Method

Abstract: This paper introduced a new Internet-based dynamic multi-document summarization system framework based on natural language processing and applied for the data management of wireless sensor networks that was capable of providing computational linguistics-related technical support. We mainly studied the related model of text mining for perceptual data. Wireless sensor networks had a large number of streaming data, with real-time characteristics. After basic node operation, it generated big data can be used for data mining. In order to apply the data mining technology to information processing of wireless sensor networks, this paper tried to find a similar such as DUC2008 abstract test samples, trained model and algorithm. The system integrated subsystems with different emphases to improve system performance, combined three innovative methods. Given that to date little research on dynamic multi-document summarization has been reported, this study had great significance. The results obtained by the new framework were compared with those from the TAC2008 evaluation, demonstrated that the new system’s performance matched that of the best existing systems.

A review on abstractive text summarization Methods

Abstract: In this article, recent key research on abstractive text summarization is reviewed. Text summarization is a mechanism of generating a brief, accurate, and precise summary of a substantial text. The most widely used strategies in text summarization are abstractive and extractive techniques. This survey is primarily concerned with abstractive text summarization and the state of the art is thoroughly discussed. The abstractive approach generates words and phrases that have not yet been used in the input corpus. Depending on the algorithm used, the approaches that are offered are divided into categories. This article includes a discussion of the various techniques applied for abstractive text summarization. This study attempts to support the necessity for text summarization, identify existing effective methods, point out problems with text summarization, and offer ways to improve performance, including potential directions for future text summarization research.

Video Summarization Based on the "Entities-Description-Utilities" Model

Abstract: In order to overcome the problem that most of the current video abstract technologies are difficultly suitable to new application environments,we present the video summarization technology based on the "Entities-Description-Utilities" model in this paper.In the model,we first use entities to produce description,then compute utility functions according to the description,and finally acquire video abstract based on the utilities.In this paper,we apply the model to testing news video summarization and get good experimental results.

Document Summarization Using Differential Evolution Algorithm

Abstract: The use of document summarization allows a user to get a sense of the content of full document, or to know its information content without reading all sentences within the document. Data reduction helps user to find the required information quickly without rendering more time in reading the whole document. This paper presents a method to generate a summary from the original document. And the method includes several characteristics such as sentence-id, position of each term in a sentence, term frequency, sentence similarity measure and weight of each and every sentence. To solve the optimization problem differential evolution (DE) algorithm is used, which can choose the optimal summary. DE algorithm is based on a fitness function and selection of fitness function is crucial for the good performance of DE algorithm.