2011
Autores
Campos, R; Dias, G; Jorge, A;
Publicação
PROGRESS IN ARTIFICIAL INTELLIGENCE
Abstract
In the last few years, a huge amount of temporal written information has become widely available on the Internet with the advent of forums, blogs and social networks. This gave rise to a new challenging problem called future retrieval, which consists of extracting future temporal information, that is known in advance, from web sources in order to answer queries that combine text of a future temporal nature. This paper aims to confirm whether web snippets can be used to form an intelligent web that can detect future expected events when their dates are already known. Moreover, the objective is to identify the nature of future texts and understand how these temporal features affect the classification and clustering of the different types of future-related texts: informative texts, scheduled texts and rumor texts. We have conducted a set of comprehensive experiments and the results show that web documents are a valuable source of future data that can be particularly useful in identifying and understanding the future temporal nature of a given implicit temporal query.
2011
Autores
Campos, R; Dias, G; Jorge, AM;
Publicação
CEUR Workshop Proceedings
Abstract
The World Wide Web (WWW) is a huge information network from which retrieving and organizing quality relevant content remains an open question for mostly all implicit temporal queries, i.e., queries without any date but with an underlying temporal intent. In this research, we aim at studying the temporal nature of any given query by means of web snippets or web query logs. For that purpose, we conducted a set of experiments, which goal is to assess the percentage of web snippets or queries (in query logs) having temporal features, thus checking whether they are a valuable source of data to help on inferring the temporal intent of queries, namely implicit ones. Our results show that web snippets, as opposed to web query logs, are an important source of concentrated information, where time clues often appear. As a consequence, they can be particularly useful to identify and understand "on-the-fly" the implicit temporal nature of queries in the context of ephemeral clustering.
2011
Autores
Brito, P; Polaillon, G;
Publicação
CEUR Workshop Proceedings
Abstract
This work comes within the field of data analysis using Galois lattices. We consider ordinal, numerical single or interval data as well as data that consist on frequency/probability distributions on a finite set of categories. Data are represented and dealt with on a common framework, by defining a generalization operator that determines intents by intervals. In the case of distribution data, the obtained concepts are more homogeneous and more easily interpretable than those obtained by using the maximum and minimum operators previously proposed. The number of obtained concepts being often rather large, and to limit the influence of atypical elements, we propose to identify stable concepts using interval distances in a cross validation-like approach.
2011
Autores
Noirhomme Fraiture, M; Brito, P;
Publicação
Statistical Analysis and Data Mining
Abstract
This paper introduces symbolic data analysis, explaining how it extends the classical data models to take into account more complete and complex information. Several examples motivate the approach, before the modeling of variables assuming new types of realizations are formally presented. Some methods for the (multivariate) analysis of symbolic data are presented and discussed. This is however far from being exhaustive, given the present dynamic development of this new field of research. Copyright © 2011 Wiley Periodicals, Inc., A Wiley Company.
2011
Autores
Prudencio, RBC; Soares, C; Ludermir, TB;
Publicação
HYBRID ARTIFICIAL INTELLIGENT SYSTEMS, PART I
Abstract
Several meta-learning approaches have been developed for the problem of algorithm selection. In this context, it is of central importance to collect a sufficient number of datasets to be used as meta-examples in order to provide reliable results. Recently, some proposals to generate datasets have addressed this issue with successful results. These proposals include datasetoids, which is a simple manipulation method to obtain new datasets from existing ones. However, the increase in the number of datasets raises another issue: in order to generate meta-examples for training, it is necessary to estimate the performance of the algorithms on the datasets. This typically requires running all candidate algorithms on all datasets, which is computationally very expensive. One approach to address this problem is the use of active learning, termed active meta-learning. In this paper we investigate the combined use of active meta-learning and datasetoids. Our results show that it is possible to significantly reduce the computational cost of generating meta-examples not only without loss of meta-learning accuracy but with potential gains.
2011
Autores
Kanda, J; Carvalho, ACPLFd; Hruschka, ER; Soares, C;
Publicação
Int. J. Hybrid Intell. Syst.
Abstract
The access to the final selection minute is only available to applicants.
Please check the confirmation e-mail of your application to obtain the access code.