XAIRE: An ensemble-based methodology for determining the relative importance of variables in regression tasks. Application to a hospital emergency department.

Nowadays it is increasingly important in many applications to understand how different factors influence a variable of interest in a predictive modeling process. This task becomes particularly important in the context of Explainable Artificial Intelligence. Knowing the relative impact of each variable on the output allows us to acquire more information about the problem and about the output provided by a model. This paper proposes a new methodology, XAIRE, that determines the relative importance of input variables in a prediction environment, considering multiple prediction models in order to increase generality and avoid bias inherent in a particular learning algorithm. Concretely, we present an ensemble-based methodology that promotes the aggregation of results from several prediction methods to obtain a relative importance ranking. Also, statistical tests are considered in the methodology in order to reveal significant differences between the relative importance of the predictor variables. As a case study, XAIRE is applied to the arrival of patients in a Hospital Emergency Department, which has resulted in one of the largest sets of different predictor variables in the literature. Results show the extracted knowledge related to the relative importance of the predictors involved in the case study.

Association of interleukin-6 polymorphisms with obesity or metabolic traits in young Mexican-Americans.

Objective: The objective of the study is to investigate the association of interleukin-6 (IL6) promoter single-nucleotide polymorphisms rs1800797 (-597 G/A) and rs1800796 (-572 G/C) with obesity or metabolic syndrome in Mexican-Americans. Methods: The rs1800797 and rs1800796 single-nucleotide polymorphisms were genotyped in Mexican-Americans (n = 437) from South Texas, and results were correlated with measures of obesity and metabolic syndrome including body mass index, waist circumference, blood pressure, cholesterol, triglycerides, glucose, liver enzymes, plasma IL6 and high-sensitive C-reactive protein (hs-CRP). Results: Significant associations were found for the rs1800796 variant with increased waist circumference, insulin resistance, lower IL6 levels and higher hs-CRP levels. The rs1800797 variant showed no associations with metabolic traits but was associated with higher IL6 levels and lower hs-CRP levels. Conclusions: Findings in this study support the anti-inflammatory, anti-obesity and glucose homeostatic roles of IL6 in Mexican-American youth.

The linear separability problem: some testing methods.

The notion of linear separability is used widely in machine learning research. Learning algorithms that use this concept to learn include neural networks (single layer perceptron and recursive deterministic perceptron), and kernel machines (support vector machines). This paper presents an overview of several of the methods for testing linear separability between two classes. The methods are divided into four groups: Those based on linear programming, those based on computational geometry, one based on neural networks, and one based on quadratic programming. The Fisher linear discriminant method is also presented. A section on the quantification of the complexity of classification problems is included.

Dendritic cell expression of ADAM23 governs T cell proliferation and cytokine production through the α(v)ß(3) integrin receptor.

ADAM23 is a member of the brain macrophage-derived chemokine family. Structural homology of ADAM proteins suggests their function as integrin receptors. Previous studies have linked ADAM23 as a dominant contributor to brain development and cancer metastasis. The present studies now show that ADAM23 expression on DCs partially governs antigen-presentation capacities to responder CD4+ T cells. With the use of RNAi approaches, knockdown of ADAM23 in murine BMDCs resulted in impaired T cell activation, proliferation, and cytokine production. Knockdown did not alter the maturation profile of DCs (i.e., costimulatory molecule expression or production of proinflammatory cytokines) but markedly impaired cognate T cell responses. There was a significant decrease in antigen-specific clonal expansion coupled with a global decrease in Th cytokine production. Impaired early activation and proliferation did not alter/skew the balance of Th polarization but significantly depressed total levels of IL-2, IFN-γ, IL-4, and IL-17 cytokine production in CD4+ T cells primed by ADAM23 knockdown versus control DCs. Finally, neutralizing antibodies targeting the α(v)ß(3) integrin receptors resulted in similar phenotypes of impaired CD4+ T cell responses. Taken together, these studies show a novel role of ADAM23 in governing DC antigen presentation to cognate CD4+ T cells.

Mechanical cardiac valve prostheses: wear characteristics and magnitudes in three bileaflet valves.

Three different bileaflet mechanical heart valves were evaluated for wear and durability characterization. The designs of the three mechanical heart valves encompass both geometrical and material interface differences. The St. Jude Medical mechanical heart valve is partially characterized by a flat-on-flat leaflet-to-orifice closing stop interface with the orifice constructed of a graphite substrate with pyrolytic carbon coating. The CarboMedicsTM mechanical heart valve is partially characterized by a flat-on-cylindrical leaflet-to-orifice closing stop interface and the orifice is constructed of solid pyrolytic carbon. The Sorin Biomedica BicarbonTM mechanical heart valve is characterized by a flat-on-flat leaflet-to-orifice closing stop interface with the leaflets constructed of pyrolytic carbon and the orifice constructed of a titanium alloy. In vitro mechanical wear analysis was performed in accordance with current FDA and ISO guidelines for accelerated life and durability testing. Comparisons revealed that the St. Jude Medical mechanical heart valve had the lowest magnitude of wear, but both the St. Jude Medical and CarboMedics mechanical heart valves proved to be very wear resistant. The Sorin Biomedica Bicarbon mechanical heart valve showed an extremely high wear rate and magnitude. The overall mechanism for material removal and wear must be fully determined to assess the long term efficacy of the mechanical heart valve.

An interlaboratory comparison of the FDA protocol for the evaluation of cavitation potential of mechanical heart valves.

Five laboratories carried out measurements of cavitation threshold for a common set of six mechanical prosthetic heart valves, two each from three different manufacturers. This study was intended to evaluate to what extent FDA's current guidance for cavitation testing would lead to consistent results in a variety of laboratory settings and to seek areas for improvement in the recommended test protocol. The inter-laboratory study protocol specified: (1) characterization of the test fluid by oxygen content and electrical conductivity, (2) location and frequency response of pressure sensors, (3) determination of ventricular and atrial pressures (P) and loading rates (dP/dt) averaged over the time period of valve closure and over the time periods of 1 ms, 5 ms, and 20 ms prior to video visualization. The protocol did not specify: (1) the fluid pumping equipment to be used to generate cavitation, (2) the pump or fluid parameters adjusted to raise or lower the loading rate, (3) the equipment, technique, or sensitivity used to visualize cavitation, and (4) a specific definition of the threshold for cavitation. Results from the five laboratories are reported. Significant differences in results were observed in dP/dt and in the pressure difference across the valves during closure at cavitation threshold. Specific differences in test systems included a wide range of ventricular compliance and single valved versus double valved test systems. Three single valve systems with compliant ventricles produced results in reasonable agreement with one another. Further similarity in test equipment should be specified to assure adequate interlaboratory reliability for cavitation testing. Areas needing better specification include the design of the valve mount, the design of the cavitation generators, and qualitative criteria for detection of threshold cavitation.

A survey of partially connected neural networks.

Almost all artificial neural networks are by default fully connected, which often implies a high redundancy and complexity. Little research has been devoted to the study of partially connected neural networks, despite its potential advantages like reduced training and recall time, improved generalization capabilities, reduced hardware requirements, as well as being a step closer to biological reality. This publication presents an extensive survey of the various kinds of partially connected neural networks, clustered into a clear framework, followed by a detailed comparative discussion.

Application of growing hierarchical SOM for visualisation of network forensics traffic data.

Digital investigation methods are becoming more and more important due to the proliferation of digital crimes and crimes involving digital evidence. Network forensics is a research area that gathers evidence by collecting and analysing network traffic data logs. This analysis can be a difficult process, especially because of the high variability of these attacks and large amount of data. Therefore, software tools that can help with these digital investigations are in great demand. In this paper, a novel approach to analysing and visualising network traffic data based on growing hierarchical self-organising maps (GHSOM) is presented. The self-organising map (SOM) has been shown to be successful for the analysis of highly-dimensional input data in data mining applications as well as for data visualisation in a more intuitive and understandable manner. However, the SOM has some problems related to its static topology and its inability to represent hierarchical relationships in the input data. The GHSOM tries to overcome these limitations by generating a hierarchical architecture that is automatically determined according to the input data and reflects the inherent hierarchical relationships among them. Moreover, the proposed GHSOM has been modified to correctly treat the qualitative features that are present in the traffic data in addition to the quantitative features. Experimental results show that this approach can be very useful for a better understanding of network traffic data, making it easier to search for evidence of attacks or anomalous behaviour in a network environment.