Catálogo de publicaciones - libros

The more complicated and large-scaled software systems get, the more important software requirements become, and detecting conflicts between requirements is one of the essential matters that must be considered for successful software projects. Formal methods have been proposed to tackle this problem by adding formality and removing ambiguity. However, they are hard to understand by non-experts, which limit their application to restricted domains. In addition, there is no overall process that covers all the steps for managing conflicts. We propose a process for systematically identifying and managing requirements conflicts. This process is composed of four steps: requirements authoring, partition, conflicts detection and conflicts management. The detection and management of the conflicts are done based on the requirements partition in natural language and supported by a tool. To demonstrate its feasibility, the proposed process has been applied to a home integration system (HIS) and the results are analyzed.

An experiment on Web-assisted detection and correction of malapropism is reported. Malapropos words semantically destroy collocations they are in, usually with retention of syntactical links with other words. A hundred English malapropisms were gathered, each supplied with its correction candidates, i.e. word combinations with one word equal to an editing variant of the corresponding word in the malapropism. Google statistics of occurrences and co-occurrences were gathered for each malapropism and correcting candidate. The collocation components may be adjacent or separated by other words in a sentence, so statistics were accumulated for the most probable distance between them. The raw Google occurrence statistics are then recalculated to numeric values of a specially defined Semantic Compatibility Index (SCI). Heuristic rules are proposed to signal malapropisms when SCI values are lower than a predetermined threshold and to retain a few highly SCI-ranked correction candidates. Within certain limitations, the experiment gave promising results.

Web Directories provide a way of locating relevant information on the Web. Typically, Web Directories rely on humans putting in significant time and effort into finding important pages on the Web and categorizing them in the Directory. In this paper we present a way for automating the creation of a Web Directory. At a high level, our method takes as input a subject hierarchy and a collection of pages. We first leverage a variety of lexical resources from the Natural Language Processing community to enrich our hierarchy. After that, we process the pages and identify sequences of important terms, which are referred to as lexical chains. Finally, we use the lexical chains in order to decide where in the enriched subject hierarchy we should assign every page. Our experimental results with real Web data show that our method is quite promising into assisting humans during page categorization.

To perform the task of email categorization, the tournament methods are proposed in this article in which the multi-class categorization process is broken down into a set of binary classification tasks. The methods of elimination tournament and Round Robin tournament are implemented and applied to classify emails within 15 folders. Substantial experiments are conducted to compare the effectiveness and robustness of the tournament methods against the n-way classification method. The experimental results prove that the tournament methods outperform the n-way method by 11.7% regarding precision, and the Round Robin performs slightly better than the Elimination tournament on average.

This paper describes the methodology, process and results of developing an application ontology as software specification of the semantics of forensics in the email suspicious of Nigerian frauds. Real life examples of fraud emails are analyzed for evidence and red flags to capture the underlying domain semantics with an application ontology of frauds. A model of the natural language structure in regular expressions is developed in the light of the ontology and applied to emails to extract linguistic evidences of frauds. The evaluation of the initial results shows a satisfactory recognition as an automatic fraud alert system. It also demonstrates a methodological significance: the methodical conceptual modeling and specific purpose-driven linguistic modeling are effective in encapsulating and managing their respective needs, perspectives and variability in real life linguistic processing applications.

This paper discusses an advanced tagging concept which supplies information that allows for a tool supported step by step mapping of natural language requirements specification to a conceptual (predesign) model.The focus lies on sentences containing conditions as used for describing alternatives in business process specifications. It is shown, how the tagging results are interpreted systematically such allowing for a stepwise model generation.

This paper studies the use of Named Entity Recognition (NER) for the Question Anwering (QA) task in Spanish texts. NER applied as a preprocessing step not only helps to detect the answer to the question but also decreases the amount of data to be considered by QA. Our proposal reduces a 26% the quantity of data and moreover increases a 9% the efficiency of the system.

It is well known that Information Retrieval Systems based entirely on syntactic contents have serious limitations. In order to achieve high precision Information Retrieval Systems the incorporation of Natural Language Processing techniques that provide semantic information is needed. For this reason, in this paper a method to determine the semantic role for the constituents of a sentence is presented. The goal of this is to integrate this method in an Information Retrieval System.

The distribution of the number of documents in topic classes is typically highly skewed. This leads to good micro-average performance but not so desirable macro-average performance. By viewing topics as clusters in a high dimensional space, we propose the use of clustering to determine subtopic clusters for large topic classes by assuming that large topic clusters are in general a mixture of a number of subtopic clusters. We used the Reuters News articles and support vector machines to evaluate whether using subtopic cluster can lead to better macro-average performance.

Successful participation in task-oriented, inference-rich dialogs requires, among other things, understanding of specifications implicitly conveyed through the exploitation of parallel structures. Several linguistic operators create specifications of this kind, including “the other way (a)round”, “vice-versa”, and “analogously”; unfortunately, automatic reconstruction of the intended specification is difficult due to the inherent dependence on given context and domain. We address this problem by a well-informed reasoning process. The techniques applied include building deep semantic representations, application of categories of patterns underlying a formal reconstruction, and using pragmatically-motivated and domain-justified preferences. Our approach is not only suitable for improving the understanding in everyday discourse, but it specifically aims at extending capabilities in a tutorial dialog system, where stressing generalities and analogies is a major concern.