![]() ![]() In the following sections, this “criterion” will be discussed in detail. Thus, such criterion could be applied to other documents and offer as a response a number of summaries that are similar to the ones that the person would have produced manually, but in less time. The goal of the method is learning the “criterion” used by a person when summarizing a set of documents. In this research work, a new method using this technique is presented, aimed at producing extractive summaries. Particle Swarm Optimization, which is the basis for the method proposed here, is a metaheuristic that, since its inception in 1995, has been successfully used in the resolution of a wide range of problems. We are constantly solving optimization problems, for instance, when we look for the fastest way to a certain location, or when we try to get things done in as little time as possible. Optimization, in the sense of finding the best solution–or at least an acceptable one–for a given problem, is still a highly significant field. However, these papers consider a set of metrics that are defined a priori, and the selection of metrics is not part of the optimization process, as for example in. Recent works consider the task of producing extractive summaries as an optimization problem, where one or more target functions are proposed to select the “best” phrases from the document to form part of the summary. Each phrase is labeled as “correct” if it is going to be part of the abstract, or “incorrect” if it is not. Obtaining such a summary can be considered as a classification problem that has two unique classes. In the case of biomedical documents, extractive is usually used. īasically, there are two ways to generate automatic summaries of texts: extractive, selecting the most relevant phrases, and abstractive, using an intermediate representation, such as a graph and verbalize it by generating new expressions in natural language. For example, the need to access and share knowledge in medicine is becoming increasingly more evident. Many texts, especially those that are non-academic or non-scientific, can be examined from different points of view and therefore different essential elements. Selecting relevant information is sometimes a more or less objective process, but, in many cases, it depends on the specific characteristics of the person summarizing the information. ![]() To summarize is to identify what is essential for a given purpose in a given context. This task, in the case of texts, is known as “summarizing”, a cognitive characteristic of human intelligence that is used to keep what is essential. We unconsciously try to retain essential information from all that information. Humans are unable to store all this information because our memory capacities are limited. More digital textual information is being consumed every day. Text documents are still the most commonly used in today’s digital society. The advances in technology achieved in recent times favored the generation of large volumes of data and, as a result, the development and application of intelligent methods capable of automating their handling have become essential. Many years after forecasting that more information than it would be possible to process would be produced, access to information and information processing became an essential need both for academics as well as for companies and organizations. The empirical results yield an improved accuracy compared to previous methods used in this field. Our paper shows that using user labeled information in the training set helps to find better metrics and weights. The proposed method combines a binary representation and a continuous one, using an original variation of the technique developed by the authors of this paper. ![]() The key feature of the proposed method is the identification of those features that are closest to the criterion used by the individual when summarizing. In this article, a new method is presented that allows automatically generating extractive summaries from documents by adequately weighting sentence scoring features using Particle Swarm Optimization. Obtaining the main contents of any given document in less time than it would take to do that manually is still an issue of interest. As information overload increases, automatic summaries allow handling the growing volume of documents, usually by assigning weights to the extracted phrases based on their significance in the expected summary. Automatic text summarization tools have a great impact on many fields, such as medicine, law, and scientific research in general. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |