Year: 2013
Type: editorial
Abstract: This issue of the IDA journal consists of ten articles which represent a variety of topics, all related to the applied and theoretical research in the field of Intelligent Data Analysis.The first three articles of this issue are about similarity and association.Khosravi-Farasani et al. in the first article discuss similarity estimation and propose a new perspective on specifying the similarity between resources in linked data and in particular for a directed and attributed graph.The idea is to use combination of structural properties of a graph and attribute value of vertices.Their suggested model is evaluated in a clustering procedure using some standard data sets where class label of each resource is estimated and compared with the ground-truth class label.Their experimental results show that their model outperforms other clustering algorithms in terms of precision and recall.The second article of this issue by Jaminéz et al. is about the study of association rules within a group of individuals.The authors define group association rules where they identify the interestingness measures for them.These measures are then used to rank not only groups of individuals but also rules within each group.The proposed approach is evaluated on a number of well-known data sets.Zare et al. in the third article of this issue discuss the drawback of computing an exact measure of betweenness centrality in high dimensional data.They apply a random projection approach and propose an approximation algorithm that is suitable for weighted and unweighted graphs.Their proposed algorithm significantly reduces the number of single-source shortest path computations.The results of their evaluation on a number of realworld networks and synthetic bench marks show interesting results based on some statistical evaluation measures.The next two articles are mostly on classification and supervised learning.Wang proposes a nonparametric bivariate copula estimation method which is based on smoothing empirical copula with shape restricted least squares support vector regression.The proposed approach exploits a priori shape of copula function.The experiments presented in the article that are based on artificial and financial time-series data, clearly show that the approach can achieve significantly better performance than parametric methods and kernel smoother.Hassanzadeh and Kevanpour in the fifth article of this article discuss the topic of sequence labeling and propose using active learning and semi-supervised learning for sequence labeling.It seems that their proposed approach could significantly reduce manual annotation costs (∼90%) compared with supervised learning that is estimated at 30%.The particular model selected here is conditional random field which has shown promising performance in several sequence labeling tasks.This issue also includes four applied research articles.Shankar and Purusothaman discuss the concept of utility based data mining and argue that when data mining is employed in stock market prediction, it would devise association rules that would lead to high return in investment.The authors propose a new analysis methodology which is based on the concept of utility based data mining where it generates utility emphasized trading rules.The article includes experiments where the proficiency of the obtained rules is compared against the conventional utility-emphasized and frequency-based trading rules using a set of related values.Mazloomian and Beigy discuss the concept of nested effects model and propose
Source: Intelligent Data Analysis
DOI: 10.3233/ida-130572
Author: A. Famili
Language: Spanish
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Keywords: similarity estimation, linked data, attributed graphs, graph clustering, interestingness measures, betweenness centrality, random projections, high-dimensional graphs, copula estimation, support vector regression, sequence labeling, active learning, semi-supervised learning, conditional random fields, stock market prediction
Open Access status: bronze
OpenAlex ID: W4242743822
