Text Summarization by Sentence Extraction Using Unsupervised Learning

García-Hernández, René Arnulfo; Montiel, Romyna; Ledeneva, Yulia; Rendón, Eréndira; Gelbukh, Alexander; Cruz, R Armas

doi:10.1007/978-3-540-88636-5_12

Cited by 27 publications

(15 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Baseline es una heurística que consiste en tomar las primeras n líneas u oraciones del texto para generar el resumen [14]. Baseline:aleatorio es una heurística cuyo funcionamiento consiste en tomar algunas oraciones del texto al azar [15]. Por lo que cualquier método que se comporte como Baseline:aleatorio no tendría razón de ser.…”

Section: Trabajo Relacionadounclassified

See 1 more Smart Citation

Evaluación de las herramientas comerciales y métodos del estado del arte para la generación de resúmenes extractivos individuales

Mendoza¹,

Ledeneva²,

García-Hernández³

et al. 2013

RCS

Self Cite

View full text Add to dashboard Cite

Resumen. Hoy en día se tiene conocimiento del avance significativo de las herramientas y métodos del estado de arte para la generación de resúmenes extractivos individuales (GREI). Para evaluar la calidad de las herramientas y los métodos de GREI es posible comparar los resúmenes generados automáticamente con los resúmenes generados por humanos. Utilizando el mismo corpus (DUC-2002) y el mismo método de evaluación independiente (ROUGE), en este trabajo se evaluaron seis herramientas comerciales y siete métodos del estado del arte que se desarrollaron para GREI. Con respecto a las herramientas comerciales de GREI, interesantemente se puede observar el avance significativo de las herramientas en línea con respecto a las herramientas instalables.Actualmente la información en formato digital se ha ido incrementando de forma exponencial, se estima que crece a un ritmo de 57% anual [1]. El incremento acelerado de la información dificulta la obtención de documentos sobre temas específicos. Por lo que sería conveniente que los documentos recuperados de una consulta en Internet fueran relevantes y no redundantes [2]. Sin embargo, normalmente no es así, por lo que es necesario contar con herramientas de procesamiento automático de texto para generar resúmenes.

show abstract

Section: Trabajo Relacionadounclassified

“…También se tienen las heurísticas Baseline [14] y Baseline:aleatorio [15] las cuales se utilizan para medir el avance que presentan los métodos del estado del arte. Sin embargo, poco se sabe de la calidad que tienen las herramientas comerciales en comparación con los métodos propuestos en el estado del arte de GREI.…”

Section: Introductionunclassified

Evaluación de las herramientas comerciales y métodos del estado del arte para la generación de resúmenes extractivos individuales

Mendoza¹,

Ledeneva²,

García-Hernández³

et al. 2013

RCS

Self Cite

View full text Add to dashboard Cite

show abstract

“…A summary obtained by extraction is composed of a set of sentences selected from the source document(s) by using statistical or heuristic methods based on information entropy of sentences. The summarization process by extraction is a relevant alternative, robust and independent of language, compared with the summarization process by abstraction [2]. An abstractive summary is obtained by semantic analysis in order to interpret the source text, and find new concepts to generate a new text that will be the summary.…”

Section: Introductionmentioning

confidence: 99%

“…The main problem of this kind automatic summarization by extraction lies in identifying the most important information of the document sources [2]. Different methods have been used until now with more or less successful results according to measurements based on recall (the number of correct sentences selected on the total number of correct sentences) and precision (the number of correct sentences on the total number of selected sentences) [6].…”

Section: Introductionmentioning

confidence: 99%

Insertion of Ontological Knowledge to Improve Automatic Summarization Extraction Methods

Motta¹,

Capus²,

Tourigny³

2011

JILSA

View full text Add to dashboard Cite

The vast availability of information sources has created a need for research on automatic summarization. Current methods perform either by extraction or abstraction. The extraction methods are interesting, because they are robust and independent of the language used. An extractive summary is obtained by selecting sentences of the original source based on information content. This selection can be automated using a classification function induced by a machine learning algorithm. This function classifies sentences into two groups: important or non-important. The important sentences then form the summary. But, the efficiency of this function directly depends on the used training set to induce it. This paper proposes an original way of optimizing this training set by inserting lexemes obtained from ontological knowledge bases. The training set optimized is reinforced by ontological knowledge. An experiment with four machine learning algorithms was made to validate this proposition. The improvement achieved is clearly significant for each of these algorithms

show abstract

“…TF-ISF is a more suitable weighting system for our system, since it ranks sentences instead of words. TFi,j is term frequency of i th index term in the j th sentence, and ISFi is inverse-sentence frequency of i th index term [98] [99]. Like the TF-IDF model, the corresponding weight for a sentence is computed as = .…”

Section: Information Assemblermentioning

confidence: 99%

Persuasive Content Generator The Design, Development and Validation of Persuasive Contect Generator Based on Social Media Profiles

Khataei¹

View full text Add to dashboard Cite

Providing content and experiences to users in order to influence their thoughts and opinions is essential in many fields such as marketing, education, and politics. With advances and the growing availability of the Internet and Web technologies, posting documents online has become an effective means of communicating to large audiences. However, given individuals' different respective interests, characteristics, and abilities, shared documents will not likely be equally persuasive to all users. To communicate persuasively, the categorization of users according to factors like age, marital status, education, and occupation is becoming increasingly prevalent, as is providing users with content specific to their categories. This approach assumes individuals within a category are similar, which is not necessarily true. Persuasive communication should recognize differences at the individual level and personalize-rather than categorize-the content.However, a systematic software solution that tailors and prepares separate persuasive content for individuals does not yet exist. Given the increasing popularity and usage of social media, socialnetwork extractable data can potentially provide a tremendous source of insight and background about individuals. Inspired by the Yale Attitude Change approach, this thesis proposes a multilayer model called Pyramid of Individualization and the related software framework to generate persuasive content based on initial author input and audience social-media data. Preliminary results show the proposed system can create personalized information that (a) matches reader interests (attention), (b) is tailored to reader ability to understand the information (comprehension), and (c) is supported by a trustable source (acceptance).ii ACKNOWLEDGMENT First and foremost, I would like to express my sincere appreciation to my supervisors Dr.Ali Arya and Dr. Michael Hine without whom this study would not have been possible. Their guidance, patience, and encouragements made all the difference.

show abstract

Text Summarization by Sentence Extraction Using Unsupervised Learning

Cited by 27 publications

References 20 publications

Evaluación de las herramientas comerciales y métodos del estado del arte para la generación de resúmenes extractivos individuales

Evaluación de las herramientas comerciales y métodos del estado del arte para la generación de resúmenes extractivos individuales

Insertion of Ontological Knowledge to Improve Automatic Summarization Extraction Methods

Persuasive Content Generator The Design, Development and Validation of Persuasive Contect Generator Based on Social Media Profiles

Contact Info

Product

Resources

About