Gene expression microarray is a rapidly maturing technology that provides the opportunity to assay the expression levels of thousands or tens of thousands of genes in a single experiment. We present a new heuristic to select relevant gene subsets in order to further use them for the classification task. Our method is based on the statistical significance of adding a gene from a ranked-list to the final subset. The efficiency and effectiveness of our technique is demonstrated through extensive comparisons with other representative heuristics. Our approach shows an excellent performance, not only at identifying relevant genes, but also with respect to the computational cost.

A groundwater treatment technology based on catalytic reductive dehalogenation has been developed to efficiently destroy chlorinated hydrocarbons in situ using a reactive well approach. The treatment process utilizes dissolved H2 as an electron donor, in the presence of a commercial palladium-on-alumina catalyst, to rapidly reduce common chlorinated aliphatics such as trichloroethylene and tetrachloroethylene into nonchlorinated hydrocarbons such as ethane. Rapid reaction rates permit the deployment of a treatment unit within a dual-screened well bore, allowing contaminated groundwater to be drawn from one water-bearing zone, treated within the well bore, and discharged to an adjacent zone with only one pass through the system. A demonstration groundwater treatment system based on this concept was evaluated in a chlorinated hydrocarbon contaminated aquifer at a major Superfund site. The system rapidly destroyed a variety of common contaminants such as TCE and PCE and maintained its performance for a test period of 1 year. Operation of the treatment system was optimized to maintain catalyst activity and to prevent formation of intermediate compounds.

No abstract

At present, automated data collection tools allow us to collect large amounts of information, not without associated problems. This paper, we apply feature selection to several software engineering databases selecting attributes with the final aim that project managers can have a better global vision of the data they manage. In this paper, we make use of attribute selection techniques in different datasets publicly available (PROMISE repository), and different data mining algorithms for classification to defect faulty modules. The results show that in general, smaller datasets with less attributes maintain or improve the prediction capability with less attributes than the original datasets.

Decision making has been traditionally based on managers experience. At present, there is a number of software engineering (SE) repositories, and furthermore, automated data collection tools allow managers to collect large amounts of information, not without associated problems. On the one hand, such a large amount of information can overload project managers. On the other hand, problems found in generic project databases, where the data is collected from different organizations, is the large disparity of its instances. In this paper, we characterize several software engineering databases selecting attributes with the final aim that project managers can have a better global vision of the data they manage. In this paper, we make use of different data mining algorithms to select attributes from the different datasets publicly available (PROMISE repository), and then, use different classifiers to defect faulty modules. The results show that in general, the smaller datasets maintain the prediction capability with a lower number of attributes than the original datasets.