Jesús S. Aguilar–Ruiz

Energy Time Series Forecasting Based on Pattern Sequence Similarity

This paper presents a new approach to forecast the behavior of time series based on similarity of pattern sequences. First, clustering techniques are used with the aim of grouping and labeling the samples from a data set. Thus, the prediction of a data point is provided as follows: first, the pattern sequence prior to the day to be predicted is extracted. Then, this sequence is searched in the historical data and the prediction is calculated by averaging all the samples immediately after the matched sequence. The main novelty is that only the labels associated with each pattern are considered to forecast the future behavior of the time series, avoiding the use of real values of the time series until the last step of the prediction process. Results from several energy time series are reported and the performance of the proposed method is compared to that of recently published techniques showing a remarkable improvement in the prediction.

Evolutionary biclustering of microarray data

In this work, we address the biclustering of gene expression data with evolutionary computation, which has been proven to have excellent performance on complex problems. In expression data analysis, the most important goal may not be finding the maximum bicluster, as it might be more interesting to find a set of genes showing similar behavior under a set of conditions. Our approach is based on evolutionary algorithms and searches for biclusters following a sequential covering strategy. In addition, we pay special attention to the fact of looking for high quality biclusters with large variation. The quality of biclusters found by our approach is discussed by means of the analysis of yeast and colon cancer datasets.

Analysis of feature rankings for classification

Different ways of contrast generated rankings by feature selection algorithms are presented in this paper, showing several possible interpretations, depending on the given approach to each study. We begin from the premise of no existence of only one ideal subset for all cases. The purpose of these kinds of algorithms is to reduce the data set to each first attributes without losing prediction against the original data set. In this paper we propose a method, feature–ranking performance, to compare different feature–ranking methods, based on the Area Under Feature Ranking Classification Performance Curve (AURC). Conclusions and trends taken from this paper propose support for the performance of learning tasks, where some ranking algorithms studied here operate.

Correlation–based scatter search for discovering biclusters from gene expression data

Scatter Search is an evolutionary method that combines existing solutions to create new offspring as the well–known genetic algorithms. This paper presents a Scatter Search with the aim of finding biclusters from gene expression data. However, biclusters with certain patterns are more interesting from a biological point of view. Therefore, the proposed Scatter Search uses a measure based on linear correlations among genes to evaluate the quality of biclusters. As it is usual in Scatter Search methodology an improvement method is included which avoids to find biclusters with negatively correlated genes. Experimental results from yeast cell cycle and human B-cell lymphoma datasets are reported showing a remarkable performance of the proposed method and measure.

A COMPARISON OF EFFORT ESTIMATION METHODS FOR 4GL PROGRAMS: EXPERIENCES WITH STATISTICS AND DATA MINING

This paper presents an empirical study analysing the relationship between a set of metrics for Fourth–Generation Languages (4GL) programs and their maintainability. An analysis has been made using historical data of several industrial projects and three different approaches: the first one relates metrics and maintainability based on techniques of descriptive statistics, and the other two are based on Data Mining techniques. A discussion on the results obtained with the three techniques is also presented, as well as a set of equations and rules for predicting the maintenance effort in this kind of programs. Finally, we have done experiments about the prediction accuracy of these methods by using new unseen data, which were not used to build the knowledge model. The results were satisfactory as the application of each technique separately provides useful perspective for the manager in order to get a complementary insight from data.

Biclustering of Gene Expression Data Based on SimUI Semantic Similarity Measure

Biclustering is an unsupervised machine learning technique that simultaneously clusters genes and conditions in gene expression data. Gene Ontology (GO) is usually used in this context to validate the biological relevance of the results. However, although the integration of biological information from different sources is one of the research directions in Bioinformatics, GO is not used in biclustering as an input data. A scatter search-based algorithm that integrates GO information during the biclustering search process is presented in this paper. SimUI is a GO semantic similarity measure that defines a distance between two genes. The algorithm optimizes a fitness function that uses SimUI to integrate the biological information stored in GO. Experimental results analyze the effect of integration of the biological information through this measure. A SimUI fitness function configuration is experimentally studied in a scatter search-based biclustering algorithm.

Shifting patterns discovery in microarrays with evolutionary algorithms

In recent years, the interest in extracting useful knowledge from gene expression data has experimented an enormous increase with the development of microarray technique. Biclustering is a recent technique that aims at extracting a subset of genes that show a similar behaviour for a subset conditions. It is important, therefore, to measure the quality of a bicluster, and a way to do that would be checking if each data submatrix follows a specific trend, represented by a pattern. In this work, we present an evolutionary algorithm for finding significant shifting patterns which depict the general behaviour within each bicluster. The empirical results we have obtained confirm the quality of our proposal, obtaining very accurate solutions for the biclusters used.

Evolutionary computation for the prediction of secondary protein structures

We have developed an evolutionary computation approach to predict secondary structure motifs using some main amino acid physical-chemical properties. The prediction model will consist of rules that predict both the beginning and the end of the regions corresponding to a secondary structure state conformation (α-helix or β-strand). A study about propensities of each pair of amino acids in capping regions of α-helix and β-strand are also performed with a data set of 12,830 non-homologous and non-redundant protein sequences.

Alpha helix prediction based on evolutionary computation

Multiple approaches have been developed in order to predict the protein secondary structure. In this paper, we propose an approach to such a problem based on evolutionary computation. The proposed approach considers various amino acids properties in order to predict the secondary structure of a protein. In particular, we will consider the hydrophobicity, the polarity and the charge of amino acids. In this study, we focus on predicting a particular kind of secondary structure: α-helices. The results of our proposal will be a set of rules that will identify the beginning or the end of such a structure.

Gene ranking from microarray data for cancer classification: a machine learning approach

Traditional gene selection methods often select the top–ranked genes according to their individual discriminative power. We propose to apply feature evaluation measure broadly used in the machine learning field and not so popular in the DNA microarray field. Besides, the application of sequential gene subset selection approaches is included. In our study, we propose some well-known criteria (filters and wrappers) to rank attributes, and a greedy search procedure combined with three subset evaluation measures. Two completely different machine learning classifiers are applied to perform the class prediction. The comparison is performed on two well–known DNA microarray data sets. We notice that most of the top-ranked genes appear in the list of relevant–informative genes detected by previous studies over these data sets.