Yulia Ledeneva

Terms derived from frequent sequences for extractive text summarization

Automatic text summarization helps the user to quickly understand large volumes of information. We present a language- and domain-independent statistical-based method for single-document extractive summarization, i.e., to produce a text summary by extracting some sentences from the given text. We show experimentally that words that are parts of bigrams that repeat more than once in the text are good terms to describe the texts contents, and so are also so-called maximal frequent sentences. We also show that the frequency of the term as term weight gives good results (while we only count the occurrences of a term in repeating bigrams).

Word Sequence Models for Single Text Summarization

The main problem for generating an extractive automatic text summary is to detect the most relevant information in the source document. For such purpose, recently some approaches have successfully employed the word sequence information from the self-text for detecting the candidate text fragments for composing the summary. In this paper, we employ the so-called n-grams and maximal frequent word sequences as features in a vector space model in order to determine the advantages and disadvantages for extractive text summarization.

Text Summarization by Sentence Extraction Using Unsupervised Learning

The main problem for generating an extractive automatic text summary is to detect the most relevant information in the source document. Although, some approaches claim being domain and language independent, they use high dependence knowledge like key-phrases or golden samples for machine-learning approaches. In this work, we propose a language- and domain-independent automatic text summarization approach by sentence extraction using an unsupervised learning algorithm. Our hypothesis is that an unsupervised algorithm can help for clustering similar ideas (sentences). Then, for composing the summary, the most representative sentence is selected from each cluster. Several experiments in the standard DUC-2002 collection show that the proposed method obtains more favorable results than other approaches.

EM clustering algorithm for automatic text summarization

Automatic text summarization has emerged as a technique for accessing only to useful information. In order to known the quality of the automatic summaries produced by a system, in DUC 2002 (Document Understanding Conference) has developed a standard human summaries called gold collection of 567 documents of single news. In this conference only five systems could outperforms the baseline heuristic in single extractive summarization task. So far, some approaches have got good results combining different strategies with language-dependent knowledge. In this paper, we present a competitive method based on an EM clustering algorithm for improving the quality of the automatic summaries using practically non language-dependent knowledge. Also, a comparison of this method with three text models is presented.

Single Extractive Text Summarization Based on a Genetic Algorithm

Extractive text summarization consists in selecting the most important units (normally sentences) from the original text, but it must be done as closer as humans do. Several interesting automatic approaches are proposed for this task, but some of them are focused on getting a better result rather than giving some assumptions about what humans use when producing a summary. In this research, not only the competitive results are obtained but also some assumptions are given about what humans tried to represent in a summary. To reach this objective a genetic algorithm is proposed with special emphasis on the fitness function which permits to contribute with some conclusions.

Comparing Commercial Tools and State-of-the-Art Methods for Generating Text Summaries

Nowadays, there are commercial tools that allow automatic generation of text summaries. However, it is not known the quality of the generated summaries and the method that it is used for the generation of the summaries using these commercial tools. This paper provides a study about the commercial tools such as Copernic Summarizer, Microsoft Office Word Summarizer 2003 and Microsoft Office Word Summarizer 2007, with the objective to detect which of them gives the summaries more similar to those made by a human. Furthermore, the comparison between commercial tools and state-of-the-art methods is realized. The experiments were carried out using DUC-2002 standard collection which contains 567 news in English.

Effect of Preprocessing on Extractive Summarization with Maximal Frequent Sequences

The task of extractive summarization consists in producing a text summary by extracting a subset of text segments, such as sentences, and concatenating them to form a summary of the original text. The selection of sentences is based on terms they contain, which can be single words or multiword expressions. In a previous work, we have suggested so-called Maximal Frequent Sequences as such terms. In this paper, we investigate the effect of preprocessing on the process of selecting such sequences. Our results suggest that the accuracy of the method is, contrary to expectations, not seriously affected by preprocessing--which is both bad and good news, as we show.

Graph Ranking on Maximal Frequent Sequences for Single Extractive Text Summarization

We suggest a new method for the task of extractive text summarization using graph-based ranking algorithms. The main idea of this paper is to rank Maximal Frequent Sequences MFS in order to identify the most important information in a text. MFS are considered as nodes of a graph in term selection step, and then are ranked in term weighting step using a graph-based algorithm. We show that the proposed method produces results superior to the-state-of-the-art methods; in addition, the best sentences were found with this method. We prove that MFS are better than other terms. Moreover, we show that the longer is MFS, the better are the results. If the stop-words are excluded, we lose the sense of MFS, and the results are worse. Other important aspect of this method is that it does not require deep linguistic knowledge, nor domain or language specific annotated corpora, which makes it highly portable to other domains, genres, and languages.

Automatic Estimation of Parameters of Complex Fuzzy Control Systems

Since the first fuzzy controller was presented by Mamdani in 1974, different studies devoted to the theory of fuzzy control have shown that the area of development of fuzzy control algorithms has been the most active area of research in the field of fuzzy logic in the last years. From 80 s, fuzzy logic has performed a vital function in the advance of practical and simple solutions for a great diversity of applications in engineering and science. Due to its great importance in navigation systems, flight control, satellite control, speed control of missiles and so on, the area of fuzzy logic has become an important integral part of industrial and manufacturing processes. Some fuzzy control applications to industrial processes have produced results superior to its equivalent obtained by classical control systems. The domain of these applications has experienced serious limitations when expanding it to more complex systems, because a complete theory does not yet exist for determining the performance of the systems when there is a change in its parameters or variables. When some of these applications are designed for more complex systems, the number of fuzzy rules controlling the process is exponentially increased with the number of variables related to the system. For example, if there are n variables and m possible linguistic labels for each variable, mn fuzzy rules would be needed to construct a complete fuzzy controller. As the number of variables n increases, the rule base quickly overloads the memory of any computing device, causing difficulties in the implementation and application of the fuzzy controller. Sensory fusion and hierarchical methods are studied in an attempt to reduce the size of the inference engine for large-scale systems. The combination of these methods reduces more considerably the number of rules than these methods separately. However, the adequate fusion-hierarchical parameters should be estimated. In traditional techniques much reliance has to be put on the experience of the system designer in order to find a good set of parameters (Jamshidi, 1997). Genetic algorithms (GA) are an appropriate technique to find parameters in a large search space. They have shown efficient and reliable results in solving optimization problems. For

A Model for Knowledge Management in Software Industry

In Micro and Small Companies for Software Development (MSCSD) of Mexico, the knowledge is generated in each software project. So, it is possible to transform tacit knowledge into explicit, to have some strategy and data storage device. Note that, when the knowledge exists only in the brains of developers there is the disadvantage that, when a developer leaves the organization, knowledge is lost. Therefore, it is possible to transform the tacit knowledge into explicit, if exist some strategy and data storage device. Developers need to generate and storage knowledge in any device and format to process, storage and exploit. In this paper, it has been called knowledge asset to an idea or solution processes or software development activities, which may be embodied as text, images, audio or video.