GenDocSum + MCLR: Generic document summarization based on maximum coverage and less redundancy

With the rapid growth of information on the Internet and electronic government recently, automatic multi-document summarization has become an important task. Multi-document summarization is an optimization problem requiring simultaneous optimization of more than one objective function. In this study, when building summaries from multiple documents, we attempt to balance two objectives, content coverage and redundancy. Our goal is to investigate three fundamental aspects of the problem, i.e. designing an optimization model, solving the optimization problem and finding the solution to the best summary. We model multi-document summarization as a Quadratic Boolean Programing (QBP) problem where the objective function is a weighted combination of the content coverage and redundancy objectives. The objective function measures the possible summaries based on the identified salient sentences and overlap information between selected sentences. An innovative aspect of our model lies in its ability to remove redundancy while selecting representative sentences. The QBP problem has been solved by using a binary differential evolution algorithm. Evaluation of the model has been performed on the DUC2002, DUC2004 and DUC2006 data sets. We have evaluated our model automatically using ROUGE toolkit and reported the significance of our results through 95% confidence intervals. The experimental results show that the optimization-based approach for document summarization is truly a promising research direction.     

Multiple documents summarization based on evolutionary optimization algorithm

This paper proposes an optimization-based model for generic document summarization. The model generates a summary by extracting salient sentences from documents. This approach uses the sentence-to-document collection, the summary-to-document collection and the sentence-to-sentence relations to select salient sentences from given document collection and reduce redundancy in the summary. To solve the optimization problem has been created an improved differential evolution algorithm. The algorithm can adjust crossover rate adaptively according to the fitness of individuals. We implemented the proposed model on multi-document summarization task. Experiments have been performed on DUC2002 and DUC2004 data sets. The experimental results provide strong evidence that the proposed optimization-based approach is a viable method for document summarization.     

MCMR: Maximum coverage and minimum redundant text summarization model

In paper, we propose an unsupervised text summarization model which generates a summary by extracting salient sentences in given document(s). In particular, we model text summarization as an integer linear programming problem. One of the advantages of this model is that it can directly discover key sentences in the given document(s) and cover the main content of the original document(s). This model also guarantees that in the summary can not be multiple sentences that convey the same information. The proposed model is quite general and can also be used for single- and multi-document summarization. We implemented our model on multi-document summarization task. Experimental results on DUC2005 and DUC2007 datasets showed that our proposed approach outperforms the baseline systems.     

Extractive multi-document summarization using population-based multicriteria optimization

Multi-document summarization is the process of extracting salient information from a set of source texts and present that information to the user in a condensed form. In this paper, we propose a multi-document summarization system which generates an extractive generic summary with maximum relevance and minimum redundancy by representing each sentence of the input document as a vector of words in Proper Noun, Noun, Verb and Adjective set. Five features, such as TF_ISF, Aggregate Cross Sentence Similarity, Title Similarity, Proper Noun and Sentence Length associated with the sentences, are extracted, and scores are assigned to sentences based on these features. Weights that can be assigned to different features may vary depending upon the nature of the document, and it is hard to discover the most appropriate weight for each feature, and this makes generation of a good summary a very tough task without human intelligence. Multi-document summarization problem is having large number of decision parameters and number of possible solutions from which most optimal summary is to be generated. Summary generated may not guarantee the essential quality and may be far from the ideal human generated summary. To address this issue, we propose a population-based multicriteria optimization method with multiple objective functions. Three objective functions are selected to determine an optimal summary, with maximum relevance, diversity, and novelty, from a global population of summaries by considering both the statistical and semantic aspects of the documents. Semantic aspects are considered by Latent Semantic Analysis (LSA) and Non Negative Matrix Factorization (NMF) techniques. Experiments have been performed on DUC 2002, DUC 2004 and DUC 2006 datasets using ROUGE tool kit. Experimental results show that our system outperforms the state of the art works in terms of Recall and Precision.     

Fuzzy evolutionary optimization modeling and its applications to unsupervised categorization and extractive summarization

Modern information retrieval (IR) systems consist of many challenging components, e.g. clustering, summarization, etc. Nowadays, without browsing the whole volume of datasets, IR systems present users with clusters of documents they are interested in, and summarize each document briefly which facilitates the task of finding the desired documents. This paper proposes a fuzzy evolutionary optimization modeling (FEOM) and its applications to unsupervised categorization and extractive summarization. In view of the nature of biological evolution, we take advantage of several fuzzy control parameters to adaptively regulate the behaviors of the evolutionary optimization, which can effectively prevent premature convergence to a local optimal solution. As a portable, modular and extensively executable model, FEOM is firstly implemented for clustering text documents. The searching capability of FEOM is exploited to explore appropriate partitions of documents such that the similarity metric of the resulting clusters is optimized. In order to further investigate its effectiveness as a generic data clustering model, FEOM is then applied to sentence clustering based extractive document summarization. It selects the most important sentence from each cluster to represent the overall meaning of document. We demonstrate the improved performance by a series of experiments using standard test sets, e.g. Reuter document collection, 20-newsgroup corpus, DUC01 and DUC02, as evaluated by some commonly used metrics, i.e. F-measure and ROUGE. The experimental results show that FEOM achieves performance as good as or better than state of arts of clustering and summarizing systems.     

An unsupervised approach to generating generic summaries of documents

We present an optimization-based unsupervised approach to automatic document summarization. In the proposed approach, text summarization is modeled as a Boolean programming problem. This model generally attempts to optimize three properties, namely, (1) relevance: summary should contain informative textual units that are relevant to the user; (2) redundancy: summaries should not contain multiple textual units that convey the same information; and (3) length: summary is bounded in length. The approach proposed in this paper is applicable to both tasks: single- and multi-document summarization. In both tasks, documents are split into sentences in preprocessing. We select some salient sentences from document(s) to generate a summary. Finally, the summary is generated by threading all the selected sentences in the order that they appear in the original document(s). We implemented our model on multi-document summarization task. When comparing our methods to several existing summarization methods on an open DUC2005 and DUC2007 data sets, we found that our method improves the summarization results significantly. This is because, first, when extracting summary sentences, this method not only focuses on the relevance scores of sentences to the whole sentence collection, but also the topic representative of sentences. Second, when generating a summary, this method also deals with the problem of repetition of information. The methods were evaluated using ROUGE-1, ROUGE-2 and ROUGE-SU4 metrics. In this paper, we also demonstrate that the summarization result depends on the similarity measure. Results of the experiment showed that combination of symmetric and asymmetric similarity measures yields better result than their use separately.     

结合LDA和谱聚类的多文档摘要

自动文摘技术的目标是致力于将冗长的文档内容压缩成较为简短的几段话,将信息全面、简洁地呈现给用户,提高用户获取信息的效率和准确率。所提出的方法在LDA（Latent Dirichlet Allocation）的基础上,使用Gibbs抽样估计主题在单词上的概率分布和句子在主题上的概率分布,结合LDA参数和谱聚类算法提取多文档摘要。该方法使用线性公式来整合句子权重,提取出字数为400字的多文档摘要。使用ROUGE自动摘要评测工具包对DUC2002数据集评测摘要质量,结果表明,该方法能有效地提高摘要的质量。     

CDDS: Constraint-driven document summarization models

This paper proposes a constraint-driven document summarization approach emphasizing the following two requirements: (1) diversity in summarization, which seeks to reduce redundancy among sentences in the summary and (2) sufficient coverage, which focuses on avoiding the loss of the document’s main information when generating the summary. The constraint-driven document summarization models with tuning the constraint parameters can drive content coverage and diversity in a summary. The models are formulated as a quadratic integer programming (QIP) problem. To solve the QIP problem we used a discrete PSO algorithm. The models are implemented on multi-document summarization task. The comparative results showed that the proposed models outperform other methods on DUC2005 and DUC2007 datasets.     

A spectral analysis approach to document summarization: Clustering and ranking sentences simultaneously

Automatic document summarization aims to create a compressed summary that preserves the main content of the original documents. It is a well-recognized fact that a document set often covers a number of topic themes with each theme represented by a cluster of highly related sentences. More important, topic themes are not equally important. The sentences in an important theme cluster are generally deemed more salient than the sentences in a trivial theme cluster. Existing clustering-based summarization approaches integrate clustering and ranking in sequence, which unavoidably ignore the interaction between them. In this paper, we propose a novel approach developed based on the spectral analysis to simultaneously clustering and ranking of sentences. Experimental results on the DUC generic summarization datasets demonstrate the improvement of the proposed approach over the other existing clustering-based approaches.     

A hybrid machine learning model for multi-document summarization

This work proposes an approach that uses statistical tools to improve content selection in multi-document automatic text summarization. The method uses a trainable summarizer, which takes into account several features: the similarity of words among sentences, the similarity of words among paragraphs, the text format, cue-phrases, a score related to the frequency of terms in the whole document, the title, sentence location and the occurrence of non-essential information. The effect of each of these sentence features on the summarization task is investigated. These features are then used in combination to construct text summarizer models based on a maximum entropy model, a naive-Bayes classifier, and a support vector machine. To produce the final summary, the three models are combined into a hybrid model that ranks the sentences in order of importance. The performance of this new method has been tested using the DUC 2002 data corpus. The effectiveness of this technique is measured using the ROUGE score, and the results are promising when compared with some existing techniques.     

TG-SMR: A Text Summarization Algorithm Based on Topic and Graph Models

Recently, automation is considered vital in most fields since computing methods have a significant role in facilitating work such as automatic text summarization. However, most of the computing methods that are used in real systems are based on graph models, which are characterized by their simplicity and stability. Thus, this paper proposes an improved extractive text summarization algorithm based on both topic and graph models. The methodology of this work consists of two stages. First, the well-known TextRank algorithm is analyzed and its shortcomings are investigated. Then, an improved method is proposed with a new computational model of sentence weights. The experimental results were carried out on standard DUC2004 and DUC2006 datasets and compared to four text summarization methods. Finally, through experiments on the DUC2004 and DUC2006 datasets, our proposed improved graph model algorithm TG-SMR (Topic Graph-Summarizer) is compared to other text summarization systems. The experimental results prove that the proposed TG-SMR algorithm achieves higher ROUGE scores. It is foreseen that the TG-SMR algorithm will open a new horizon that concerns the performance of ROUGE evaluation indicators.     

Extractive single-document summarization based on genetic operators and guided local search

Due to the exponential growth of textual information available on the Web, end users need to be able to access information in summary form – and without losing the most important information in the document when generating the summaries. Automatic generation of extractive summaries from a single document has traditionally been given the task of extracting the most relevant sentences from the original document. The methods employed generally allocate a score to each sentence in the document, taking into account certain features. The most relevant sentences are then selected, according to the score obtained for each sentence. These features include the position of the sentence in the document, its similarity to the title, the sentence length, and the frequency of the terms in the sentence. However, it has still not been possible to achieve a quality of summary that matches that performed by humans and therefore methods continue to be brought forward that aim to improve on the results. This paper addresses the generation of extractive summaries from a single document as a binary optimization problem where the quality (fitness) of the solutions is based on the weighting of individual statistical features of each sentence – such as position, sentence length and the relationship of the summary to the title, combined with group features of similarity between candidate sentences in the summary and the original document, and among the candidate sentences of the summary. This paper proposes a method of extractive single-document summarization based on genetic operators and guided local search, called MA-SingleDocSum. A memetic algorithm is used to integrate the own-population-based search of evolutionary algorithms with a guided local search strategy. The proposed method was compared with the state of the art methods UnifiedRank, DE, FEOM, NetSum, CRF, QCS, SVM, and Manifold Ranking, using ROUGE measures on the datasets DUC2001 and DUC2002. The results showed that MA-SingleDocSum outperforms the state of the art methods.     

Graph Ranked Clustering Based Biomedical Text Summarization Using Top k Similarity

Text Summarization models facilitate biomedical clinicians and researchers in acquiring informative data from enormous domain-specific literature within less time and effort. Evaluating and selecting the most informative sentences from biomedical articles is always challenging. This study aims to develop a dual-mode biomedical text summarization model to achieve enhanced coverage and information. The research also includes checking the fitment of appropriate graph ranking techniques for improved performance of the summarization model. The input biomedical text is mapped as a graph where meaningful sentences are evaluated as the central node and the critical associations between them. The proposed framework utilizes the top k similarity technique in a combination of UMLS and a sampled probability-based clustering method which aids in unearthing relevant meanings of the biomedical domain-specific word vectors and finding the best possible associations between crucial sentences. The quality of the framework is assessed via different parameters like information retention, coverage, readability, cohesion, and ROUGE scores in clustering and non-clustering modes. The significant benefits of the suggested technique are capturing crucial biomedical information with increased coverage and reasonable memory consumption. The configurable settings of combined parameters reduce execution time, enhance memory utilization, and extract relevant information outperforming other biomedical baseline models. An improvement of 17% is achieved when the proposed model is checked against similar biomedical text summarizers.     

A novel approach to update summarization using evolutionary manifold-ranking and spectral clustering

Update summarization is a new challenge in automatic text summarization. Different from the traditional static summarization, it deals with the dynamically evolving document collections of a single topic changing over time, which aims to incrementally deliver salient and novel information to a user who has already read the previous documents. How to have a content selection and linguistic quality control in a temporal context are the two new challenges brought by update summarization. In this paper, we address a novel content selection framework based on evolutionary manifold-ranking and normalized spectral clustering. The proposed evolutionary manifold-ranking aims to capture the temporal characteristics and relay propagation of information in dynamic data stream and user need. This approach tries to keep the summary content to be important, novel and relevant to the topic. Incorporation with normalized spectral clustering is to make summary content have a high coverage for each sub-topic. Ordering sub-topics and selecting sentences are dependent on the rank score from evolutionary manifold-ranking and the proposed redundancy removal strategy with exponent decay. The evaluation results on the update summarization task of Text Analysis Conference (TAC) 2008 demonstrate that our proposed approach is competitive. In the 71 run systems, we receive three top 1 under PYRAMID metrics, ranking 13th in ROUGE-2, 15th in ROUGE-SU4 and 21st in BE.     

基于篇章主次关系的单文档抽取式摘要方法研究

抽取式自动文摘研究抽取文档中最能代表文档核心内容的句子作为摘要,篇章主次关系分析则是从篇章结构方面分析出篇章的主要内容和次要内容,因此,篇章主次关系分析和抽取式自动文摘存在较大关联,篇章主次关系可指导摘要的抽取。该文提出了一种基于篇章主次关系的单文档抽取式摘要方法,该方法基于神经网络模型构建了一个篇章主次关系和文本摘要联合学习的模型。该模型在考虑词组、短语等语义信息的基础上同时考虑了篇章的主次关系等结构信息,最终基于篇章内容的整体优化抽取出最能代表文档核心内容的句子作为摘要。实验结果表明,与当前主流的单文档抽取式摘要方法相比,该方法在ROUGE评价指标上有显著提高。     

SumCR: A new subtopic-based extractive approach for text summarization

In text summarization, relevance and coverage are two main criteria that decide the quality of a summary. In this paper, we propose a new multi-document summarization approach SumCR via sentence extraction. A novel feature called Exemplar is introduced to help to simultaneously deal with these two concerns during sentence ranking. Unlike conventional ways where the relevance value of each sentence is calculated based on the whole collection of sentences, the Exemplar value of each sentence in SumCR is obtained within a subset of similar sentences. A fuzzy medoid-based clustering approach is used to produce sentence clusters or subsets where each of them corresponds to a subtopic of the related topic. Such kind of subtopic-based feature captures the relevance of each sentence within different subtopics and thus enhances the chance of SumCR to produce a summary with a wider coverage and less redundancy. Another feature we incorporate in SumCR is Position, i.e., the position of each sentence appeared in the corresponding document. The final score of each sentence is a combination of the subtopic-level feature Exemplar and the document-level feature Position. Experimental studies on DUC benchmark data show the good performance of SumCR and its potential in summarization tasks.     

The two-stage unsupervised approach to multidocument summarization

This paper suggests an approach for creating a summary for a set of documents with revealing the topics and extracting informative sentences. The topics are determined through clustering of sentences, and the informative sentences are extracted using the ranking algorithm. The result of the summarization has been shown depends on the clustering method, the ranking algorithm, and the similarity measure. The experiments on an open benchmark datasets DUC2001 and DUC2002 have showed that the suggested clustering methods and the ranking algorithm show better results than the known k-means method and the ranking algorithms PageRank and HITS.     

Automatic text summarization using latent semantic analysis

In the paper, the most state-of-the-art methods of automatic text summarization, which build summaries in the form of generic extracts, are considered. The original text is represented in the form of a numerical matrix. Matrix columns correspond to text sentences, and each sentence is represented in the form of a vector in the term space. Further, latent semantic analysis is applied to the matrix obtained to construct sentences representation in the topic space. The dimensionality of the topic space is much less than the dimensionality of the initial term space. The choice of the most important sentences is carried out on the basis of sentences representation in the topic space. The number of important sentences is defined by the length of the demanded summary. This paper also presents a new generic text summarization method that uses nonnegative matrix factorization to estimate sentence relevance. Proposed sentence relevance estimation is based on normalization of topic space and further weighting of each topic using sentences representation in topic space. The proposed method shows better summarization quality and performance than state-of-the-art methods on the DUC 2001 and DUC 2002 standard data sets.     

一种基于演化算法进行句子抽取的多文档自动摘要系统SBGA

SBGA系统将多文档自动摘要过程视为一个从源文档集中抽取句子的组合优化过程,并用演化算法来求得近似最优解。与基于聚类的句子抽取方法相比,基于演化算法进行句子抽取的方法是面向摘要整体的,因此能获得更好的近似最优摘要。演化算法的评价函数中考虑了衡量摘要的4个标准:长度符合用户要求、信息覆盖率高、更多地保留原文传递的重要信息、无冗余。另外,为了提高词频计算的精度, SBGA采用了一种改进的词频计算方法TFS,将加权后词的同义词频率加到了原词频中。在DUC2004测试数据集上的实验结果表明,基于演化算法进行句子抽取的方法有很好的性能,其ROUGE-1分值比DUC2004最优参赛系统仅低0.55%。改进的词频计算方法TFS对提高文档质量也起到了良好的作用。     

基于事件项语义图聚类的多文档摘要方法

基于事件的抽取式摘要方法一般首先抽取那些描述重要事件的句子,然后把它们重组并生成摘要。该文将事件定义为事件项以及与其关联的命名实体,并聚焦从外部语义资源获取的事件项语义关系。首先基于事件项语义关系创建事件项语义关系图并使用改进的DBSCAN算法对事件项进行聚类,接着为每类选择一个代表事件项或者选择一类事件项来表示文档集的主题,最后从文档抽取那些包含代表项并且最重要的句子生成摘要。该文的实验结果证明在多文档自动摘要中考虑事件项语义关系是必要的和可行的。