Important Sentences Research Articles

A Graph-based Approach of Automatic Keyphrase Extraction

Existing graph-based ranking techniques for keyphrase extraction only consider the connections between words in a document, ignoring the impact of the sentence. Motivated by the fact that a word must be important if it appears in many important sentences, we propose to take full advantage of the reinforcement between words andsentences by melting three kinds of relationships between them. Moreover, a document is grouped with many topics. The extracted keyphrases should be synthetic in the sense that they should deal with all the main topics in a document. Inspired by this, we take topic model into consider. Experimental results show that our approach performs betterthan state-of-the-art keyphrase extraction method on two datasets under three evaluation metrics.

Enabling Semantic Search Based on Conceptual Graphs over Encrypted Outsourced Data

Currently, searchable encryption is a hot topic in the field of cloud computing. The existing achievements are mainly focused on keyword-based search schemes, and almost all of them depend on predefined keywords extracted in the phases of index construction and query. However, keyword-based search schemes ignore the semantic representation information of users’ retrieval and cannot completely match users’ search intention. Therefore, how to design a content-based search scheme and make semantic search more effective and context-aware is a difficult challenge. In this paper, for the first time, we define and solve the problems of semantic search based on conceptual graphs (CGs) over encrypted outsourced data in clouding computing (SSCG). We first employ the efficient measure of “sentence scoring” in text summarization and Tregex to extract the most important and simplified topic sentences from documents. We then convert these simplified sentences into CGs. To perform quantitative calculation of CGs, we design a new method that can map CGs to vectors. Next, we rank the returned results based on “text summarization score”. Furthermore, we propose a basic idea for SSCG and give a significantly improved scheme to satisfy the security guarantee of searchable symmetric encryption (SSE). Finally, we choose a real-world dataset, i.e., the CNN dataset to test our scheme. The results obtained from the experiment show the effectiveness of our proposed scheme.

Automatic Summarization considering Thread Structure and Time Series in Electronic Bulletin Board System for Discussion

On electronic bulletin board systems for discussion, a topic the users argue diversifies into multiple subtopics, and the entire structure becomes complicated. It is helpful to show users summarizations of the arguments because they can help in understanding the con- tents more easily without looking over from beginning to end of the discussion forum. The purpose of this paper is to propose an automatic summarization method of a single thread considering time series, reply relationships and user information. In the proposed method, a thread is restructured in several clusters by hierarchical clustering, and important sentences compressed with linguistic relationship of predicate argument structures are selected within each cluster using LexRank, which is a stochastic graph-based method for computing the relative importance of textual units. Finally, we conducted quantitative and qualitative anal- ysis, comparing the proposed method with MMR. Both experimental results demonstrate that the proposed method can reduce redundancies more and extract fewer sentences unre- lated to the whole context of the summary than the baseline. However, the proposed method included fewer important words than the baseline.

A Graph Based Conceptual Mining Model for Abstractive Text Summarization

Objectives: Themain objective of automatic text summarization is to compress the document into a smaller version by preserving the important concepts. Methods/Statistical Analysis:This work proposes a hybrid approach of Singular Value Decomposition and Named Entity Recognition to extract important sentences present in a document. The extracted sentences are used to create a probabilistic graphical model calledaBeliefcNetwork. This graph model represents documentsummary in concept level. We have used a modified Page Rank algorithm to find the most ranked noun phrase. From this noun phrase we extracted the most relevant sentences. Findings: Our abstractive graph based model for a document generates novel sentences as it uses the concept of triplets (Subject, Verb, and Object). It identifies whether a sentence is created by structural rearrangement of another sentence.Using SVD(Singualr Value Decomposition) and NER(Named Entity Recognition) we extracted relevant information present in a document so that entire document is crushed in to a graph model. We can use this model for documents similarity as well as for plagiarism detection. Application/Improvements.Experimental results of our proposed system show that use of named entities and SVD increases the accuracy of summarizer.

Question Word in the Mandarin Language

In an interrogative sentence in Mandarin language, a question word can be placed in the beginning, middle or end of a sentence. Because of the different nationand culture, when a foreign student learns Mandarin, they find it difficult to understandthe question words and the position of the question words in that language. Because ofthat, the writer proposes to explain such problems. This research aims to find out whatare the types of question words in Mandarin, and also to explain the function and usageof question words in the Mandarin interrogative sentence. An interrogative sentence isa very important sentence. In Mandarin, the following question words: ?(shu) Who????(zi n?li) where, ???(zi n?er) where????(wi shnme) why, ??(z?nme) why????du? sh?o? how many???(du? ji?) how long????? (shnme shhu) when???(shnme) what????(zu shnme) why?? ??(gn shnme) why???(gnma) why and so on are used to ask who, where, what, how much, when, what time, and why. Those words havedifferent functions and usage.Each sentence has a certain structure and word order. A question word can beplaced in the beginning, middle, or end of a sentence. When the place is changed, thereis a possibility of miscommunication.DOI: https://doi.org/10.24071/llt.2013.160106

Improving Readability through Individualized Summary Extraction, Using Interactive Genetic Algorithm

ABSTRACTThe impact of individualized summarization is high, because a summary would be difficult to understand by all if summarized in a generic manner. When sentences are important as well as readable to the learner with reading difficulties, the summary might be useful for better comprehension of the text. In this article, we propose an Interactive Genetic Algorithm-based individualized summarization to maximize the readability of selected important sentences. The individualized summarization applies to the educational domain using readability-based features to extract readable important sentences. Inclusion of features representing reading difficulty should not dilute the informative score of the summary, moreover it aids the learner who has reading difficulties to comprehend the complete text better, using the summary as supplementary to the complete text and not as a substitute for it. The experimental results derived from intrinsic evaluation shows that individual summary extraction performs better than a GA-based approach and a baseline approach. Moreover, user -based direct evaluation also supports individualized summarization for improving the readability by the target audience.

AUTOMATIC TEXT SUMMARIZATION USING SUPERVISED MACHINE LEARNING TECHNIQUE FOR HINDI LANGAUGE

Automatic text summarization is a technique which compresses large text into a shorter text which includes the important information. Hindi is the top-most language used in India and also in a few neighboring countries there is a lack of proper summarization system for Hindi text. Hence,in this paper, we present an approach to the design an automatic text summarizer for Hindi text that generates a summary by extracting sentences. It deals with a single document summarization based on machine learning approach. Each sentence in the document is represented by a set of various features namelysentence paragraph position, sentence overall position, numeric data, presence of inverted commas, sentence length and keywords in sentences. The sentences are classified into one of four classes namelymost important, important, less important and not important. The classes are in turn having ranks from 4 to 1 respectively with “4”indicating most important sentence and “1” being least relevant sentence . Next a supervised machine learning tool SVM rank is used to train the summarizer to extract important sentences, based on the feature vector. The sentences are ordered according to the ranking of classes. Then based on the required compression ratio, sentences are included in the final summary. The experiment was performed on news articles of different category such as bollywood, politics and sports. The performance of the technique is compared with the human generated summaries. The average result of experiments indicates 72% accuracy at 50% compression ratio and 60% accuracy at 25% compression ratio.

Abstractive Summarization

Customization of information from web documents is an immense job that involves mainly the shortening of original texts. This task is carried out using summarization techniques. In general, an automatically generated summary is of two types – extractive and abstractive. Extractive methods use surface level and statistical features for the selection of important sentences, without considering the meaning conveyed by those sentences. In contrast, abstractive methods need a formal semantic representation, where the selection of important components and the rephrasing of the selected components are carried out using the semantic features associated with the words as well as the context. Furthermore, a deep linguistic analysis is needed for generating summaries. However, the bottleneck behind abstractive summarization is that it requires semantic representation, inference rules and natural language generation. In this paper, The authors propose a semi-supervised bootstrapping approach for the identification of important components for abstractive summarization. The input to the proposed approach is a fully connected semantic graph of a document, where the semantic graphs are constructed for sentences, which are then connected by synonym concepts and co-referring entities to form a complete semantic graph. The direction of the traversal of nodes is determined by a modified spreading activation algorithm, where the importance of the nodes and edges are decided, based on the node and its connected edges under consideration. Summary obtained using the proposed approach is compared with extractive and template based summaries, and also evaluated using ROUGE scores.

Improving Multi-Document Summary Method Based on Sentence Distribution

Automatic multi-document summaries had been developed by researchers. The method used to select sentences from the source document would determine the quality of the summary result. One of the most popular methods used in weighting sentences was by calculating the frequency of occurrence of words forming the sentences. However, choosing sentences with that method could lead to a chosen sentence which didn't represent the content of the source document optimally. This was because the weighting of sentences was only measured by using the number of occurrences of words. This study proposed a new strategy of weighting sentences based on sentences distribution to choose the most important sentences which paid much attention to the elements of sentences that were formed as a distribution of words. This method of sentence distribution enables the extraction of an important sentence in multi-document summarization which served as a strategy to improve the quality of sentence summaries. In that respect were three concepts used in this study: (1) clustering sentences with similarity based histogram clustering, (2) ordering cluster by cluster importance and (3) selection of important sentence by sentence distribution. Results of experiments showed that the proposed method had a better performance when compared with SIDeKiCK and LIGI methods. Results of ROUGE-1 showed the proposed method increasing 3% compared with the SIDeKiCK method and increasing 5.1% compared with LIGI method. Results of ROUGE-2 proposed method increase 13.7% compared with the SIDeKiCK and increase 14.4% compared with LIGI method.

A Summarizer for Tamil Language Using Centroid Approach

Document summarization plays a vital role in the use and management of information dissemination. This paper investigates a method for the production of summaries from Tamil newspaper text document. The primary goal is to create an effective and efficient tool that is able to summarize the given text documents in a form of meaningful extract of the original text document using centroid-based algorithm. The paper focuses on generating summaries using a centroid-based algorithm, which represents group of words that are statistically important for a document. Each sentence in a document is considered as a vector in a multi-dimensional space. The sentences that are nearest to the centroid value are considered as the most important sentences. The importance of a sentence is determined by three parameters the centroid value, the positional value, and the first sentence overlap. The score for each sentence is calculated and the redundancy between the sentences is eliminated using CSIS. Finally, the sentences are ranked and the sentences with highest score values are selected as summary.

Extractive Text Summarization System using Fuzzy Clustering Algorithm

The digital world is full of tons and tons of electronic documents. This kind of large amount of information is cumbersome to read at a glance and also it takes more time to download the large content. In this paper, we propose sentence ranking measure to process the sentences and rank it with appropriate sentence rank score. Similar sentences are grouped into clusters. Based on the score, the sentences with highest score are considered as important sentences. The sentences with top scores from each cluster are retrieved for summary report. This summary report eases the user to quickly analyse the information. The experimental result proves the effectiveness of the system.

A tracking and summarization system for online Chinese news topics

Purpose – The purpose of this paper is to design and implement new tracking and summarization algorithms for Chinese news content. Based on the proposed methods and algorithms, the authors extract the important sentences that are contained in topic stories and list those sentences according to timestamp order to ensure ease of understanding and to visualize multiple news stories on a single screen. Design/methodology/approach – This paper encompasses an investigational approach that implements a new Dynamic Centroid Summarization algorithm in addition to a Term Frequency (TF)-Density algorithm to empirically compute three target parameters, i.e., recall, precision, and F-measure. Findings – The proposed TF-Density algorithm is implemented and compared with the well-known algorithms Term Frequency-Inverse Word Frequency (TF-IWF) and Term Frequency-Inverse Document Frequency (TF-IDF). Three test data sets are configured from Chinese news web sites for use during the investigation, and two important findings are obtained that help the authors provide more precision and efficiency when recognizing the important words in the text. First, the authors evaluate three topic tracking algorithms, i.e., TF-Density, TF-IDF, and TF-IWF, with the said target parameters and find that the recall, precision, and F-measure of the proposed TF-Density algorithm is better than those of the TF-IWF and TF-IDF algorithms. In the context of the second finding, the authors implement a blind test approach to obtain the results of topic summarizations and find that the proposed Dynamic Centroid Summarization process can more accurately select topic sentences than the LexRank process. Research limitations/implications – The results show that the tracking and summarization algorithms for news topics can provide more precise and convenient results for users tracking the news. The analysis and implications are limited to Chinese news content from Chinese news web sites such as Apple Library, UDN, and well-known portals like Yahoo and Google. Originality/value – The research provides an empirical analysis of Chinese news content through the proposed TF-Density and Dynamic Centroid Summarization algorithms. It focusses on improving the means of summarizing a set of news stories to appear for browsing on a single screen and carries implications for innovative word measurements in practice.

Towards an Efficient Approach for Automatic Medical Document Summarization

Abstract Document summarization deals with providing condensed version of the original document. We present an extractive informative single medical document summarization approach. We compare the tokens in the sentence with cue words. A sentence ranking method is used to extract the important sentences. The existing summarizers are used for performance analysis.

Summary generation approaches based on semantic analysis for news documents

With the exponential growth of the internet, a lot of online news reports are produced on the web every day. The news stream flows so rapidly that no one has the time to look at each and every item of information. In this situation, a person would naturally prefer to read updated information at certain time intervals. Document updating technique is very helpful for individuals to acquire new information or knowledge by eliminating out-of-date or redundant information. Existing summarization systems involve identifying the most relevant sentences from the text and putting them together to create a concise initial summary. In the process of identifying the important sentences, features influencing the relevance of sentences are determined. Based on these features the salience of the sentence is calculated and an initial summary is generated from highly important sentences at different compression rates. These types of initial summaries work on a batch of documents and do not consider the documents that may arrive at later time, so that corresponding summaries need to get updated. The update summarization system addresses this issue by taking into account the documents read by the user in the past and seeks to present only fresh or different information. The first step is to create an initial summary based on basic and additional features. The next step is to create an update summary based on the basic, additional and update features. In this paper, two approaches are proposed for generating initial and update summary from multiple documents about given news. The first approach performs semantic analysis by modifying the vector space model with dependency parse relations and applying latent semantic analysis on it to create a summary. The second approach applies sentence annotation based on aspects, prepositions and named entities to generate summary. Experimental results show that the proposed approaches generate better initial and update summaries compared with the existing systems.

Automatic Text Summarization UsingRegression Model (GA)

Text Summarization provide large text data into a shorter version without changing its information content and meaning. It is very difficult for human beings to read whole document to understand and manually summarize the documents. Text Summarization methods are divide into two types: extractive and abstractive summarization. Anabstractive summarization is understanding of document, finding the new concepts and providing summary in few words (sentences different form texts sentences).it is very hard or impossible to design (now a days). In Extractive summarization method select important sentences, paragraphs etc. from the original text document and concatenating them into shorter form. The importance of sentences is decided based on some features of sentences. In this Research paper, Automatic Text Summarization using Extractive techniques with the help of Genetic Algorithm has been presented.

Extractive Broadcast News Summarization Leveraging Recurrent Neural Network Language Modeling Techniques

Extractive text or speech summarization manages to select a set of salient sentences from an original document and concatenate them to form a summary, enabling users to better browse through and understand the content of the document. A recent stream of research on extractive summarization is to employ the language modeling (LM) approach for important sentence selection, which has proven to be effective for performing speech summarization in an unsupervised fashion. However, one of the major challenges facing the LM approach is how to formulate the sentence models and accurately estimate their parameters for each sentence in the document to be summarized. In view of this, our work in this paper explores a novel use of recurrent neural network language modeling (RNNLM) framework for extractive broadcast news summarization. On top of such a framework, the deduced sentence models are able to render not only word usage cues but also long-span structural information of word co-occurrence relationships within broadcast news documents, getting around the need for the strict bag-of-words assumption. Furthermore, different model complexities and combinations are extensively analyzed and compared. Experimental results demonstrate the performance merits of our summarization methods when compared to several well-studied state-of-the-art unsupervised methods.

Extractive Summarization Technique Based on Fuzzy Membership Calculation Using Rough Sets

Selection of important sentences from a document is an important task in Automatic text summarization systems. When these sentences are presented without any modification in their syntax and semantics, they form extractive summary of the given text document. Extractive summarization techniques based on statistical methods like word frequency, indicator phrases and word co-occurrence are language independent. Recent studies have shown that word co-occurrence information can improve the quality of extractive summaries. This paper presents a statistical technique to identify important sentences of a text document using word co-occurrence information. The proposed technique adopts the rough set approach of Reduct and Core calculation for a given data. Fuzzy core concept is used, when the Core is empty or whenever the Core has to be enhanced. The proposed technique is tested with DUC 2002 data Sets and has given good results.

Combining Relevance Language Modeling and Clarity Measure for Extractive Speech Summarization

Extractive speech summarization, which purports to select an indicative set of sentences from a spoken document so as to succinctly represent the most important aspects of the document, has garnered much research over the years. In this paper, we cast extractive speech summarization as an ad-hoc information retrieval (IR) problem and investigate various language modeling (LM) methods for important sentence selection. The main contributions of this paper are four-fold. First, we explore a novel sentence modeling paradigm built on top of the notion of relevance, where the relationship between a candidate summary sentence and a spoken document to be summarized is discovered through different granularities of context for relevance modeling. Second, not only lexical but also topical cues inherent in the spoken document are exploited for sentence modeling. Third, we propose a novel clarity measure for use in important sentence selection, which can help quantify the thematic specificity of each individual sentence that is deemed to be a crucial indicator orthogonal to the relevance measure provided by the LM-based methods. Fourth, in an attempt to lessen summarization performance degradation caused by imperfect speech recognition, we investigate making use of different levels of index features for LM-based sentence modeling, including words, subword-level units, and their combination. Experiments on broadcast news summarization seem to demonstrate the performance merits of our methods when compared to several existing well-developed and/or state-of-the-art methods.

Automatic Text Summarization

This paper investigates on sentence extraction based single Document summarization. It saves time in our daily work once we get summarized data. Today there are so many Documents, articles, papers and reports available in digital form, but most of them lack summaries. Automatic text Summarization is a technique where a computer summarizes a text. A text is given to the computer and the computer returns a required extract of the original text document. Our methods on the sentence extraction-based text summarization task use the graph based algorithm to calculate importance of each sentence in document and most important sentences are extracted to generate document summary. These extraction based text summarization methods give an indexing weight to the document terms to compute the similarity values between sentences

Blending Sentence Optimization Weights of Unsupervised Approaches for Extractive Speech Summarization

Abstract This paper evaluates the performance of two unsupervised approaches, Maximum Marginal Relevance (MMR) and concept-based global optimization framework for speech summarization. Automatic summarization is very useful techniques that can help the users browse a large amount of data. This study focuses on automatic extractive summarization on multi-dialogue speech corpus. We propose improved methods by blending each unsupervised approach at sentence level. Sentence level information is leveraged to improve the linguistic quality of selected summaries. First, these scores are used to filter sentences for concept extraction and concept weight computation. Second, we pre-select a subset of candidate summary sentences according to their sentence weights. Last, we extend the optimization function to a joint optimization of concept and sentence weights to cover both important concepts and sentences. Our experimental results show that these methods can improve the system performance comparing to the concept-based optimization baseline for both human transcripts and ASR output. The best scores are achieved by combining all three approaches, which are significantly better than the baseline system.