Year: 2016
Type: editorial
Abstract: This issue of the IDA journal, the fifth issue of 2016, consists of twelve articles, all covering a wide range of topics related to the theoretical and applied research in the field of Intelligent Data Analysis.The first three articles of this issue are about various aspects of data preprocessing in IDA.Zhang et al. in the first article of this issue discuss environmental sensing using multitudes of wirelessly connected sensors where commodity sensors are installed and the data are known to be unreliable and noisy.The authors propose a sensor reliability-based cleaning method, called Inuence Mean (IM), which uses weighted aggregation based on individual sensor reliabilities.Their experimental results show that the method generally improves the data cleaning accuracy, particularly when the behaviors of unreliable sensors vary drastically from reliable sensors.Nemtsov et al. in the second article of this issue discuss matrix compression using the Nystrom method where they extend the applicability of this method and describe how it can be applied to find the singular value decomposition (SVD) of general matrices and the eigenvalue decomposition (EVD) of square matrices.There are three contributions to their proposed method: first, it allows the compression of a general matrix M; second, it allows the approximation of the SVD and EVD when they cannot be directly calculated due to space and time limitations; and third, a novel algorithm for selecting the initial sample is presented.Salama et al. in the third article of this group discuss the topic of data reduction for classification using ant colony algorithms, and present a new ant colony optimization (ACO) algorithm for data reduction which is based on both feature and instance selection.The approach is intended to improve the predictive quality of the constructed classification models.Their empirical evaluations on a large number of benchmark datasets with five well-known classification algorithms show that their approach improves the predictive quality of the produced classifiers.The next group of articles are about various forms on learning.Vagner, in the first article of this group, introduces the OPTICS algorithm which is a hierarchical density-based clustering method.The proposed algorithm builds a grid structure to reduce the number of data points and in order to get the clusters, the algorithm uses the reachability plots of the grid structure to determine to which cluster the original input points belong.The experimental results presented in the paper show that the proposed approach is fast and the speed-up can be one or two orders of magnitude or more, depending mainly on certain parameter of the algorithm.Wang et al. in the fifth article of this issue, discuss ensemble learning via manipulating the training set and present a strategy which is intended to combine learning set resampling and random subspace method applied to high-dimensional domains.The authors propose a new procedure, called Bag of Little Bootstraps on Features (BLBF), which works by combining the results of bootstrapping multiple feature subsets of the original dataset using the random subspace method.Their empirical experiments on various high-dimensional datasets demonstrate that the proposed approach outperforms the state-of-the-art instance-based resampling learning algorithms.Ferber-Hernandez et al. in the next article propose a Sequential-Patterns Classifier, which is based on a novel pruning strategy, using the Netconf as measure of interest, that allows to prune the rules search space for building specific rules with high Netconf.The proposed classifier was evaluated using Weka, a popular suite of machine learning software.The experiments reported were conducted using several document collections.
Source: Intelligent Data Analysis
DOI: 10.3233/ida-160852
Language: Spanish
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4233556230
