Unified framework for construction of rule based classification systems

Automatic generation of classification rules has been an increasingly popular technique in commercial applications such as Big Data analytics, rule based expert systems and decision making systems. However, a principal problem that arises with most methods for generation of classification rules is the overfit-ting of training data. When Big Data is dealt with, this may result in the generation of a large number of complex rules. This may not only increase computational cost but also lower the accuracy in predicting further unseen instances. This has led to the necessity of developing pruning methods for the simplification of rules. In addition, classification rules are used further to make predictions after the completion of their generation. As efficiency is concerned, it is expected to find the first rule that fires as soon as possible by searching through a rule set. Thus a suit-able structure is required to represent the rule set effectively. In this chapter, the authors introduce a unified framework for construction of rule based classification systems consisting of three operations on Big Data: rule generation, rule simplification and rule representation. The authors also review some existing methods and techniques used for each of the three operations and highlight their limitations. They introduce some novel methods and techniques developed by them recently. These methods and techniques are also discussed in comparison to existing ones with respect to efficient processing of Big Data.

The identification of genes important in pseudomonas syringae pv. phaseolicola plant colonisation using in vitro screening of transposon libraries

The bacterial plant pathogen Pseudomonas syringae pv. phaseolicola (Pph) colonises the surface of common bean plants before moving into the interior of plant tissue, via wounds and stomata. In the intercellular spaces the pathogen proliferates in the apoplastic fluid and forms microcolonies (biofilms) around plant cells. If the pathogen can suppress the plant’s natural resistance response, it will cause halo blight disease. The process of resistance suppression is fairly well understood, but the mechanisms used by the pathogen in colonisation are less clear. We hypothesised that we could apply in vitro genetic screens to look for changes in motility, colony formation, and adhesion, which are proxies for infection, microcolony formation and cell adhesion. We made transposon (Tn) mutant libraries of Pph strains 1448A and 1302A and found 106/1920 mutants exhibited alterations in colony morphology, motility and biofilm formation. Identification of the insertion point of the Tn identified within the genome highlighted, as expected, a number of altered motility mutants bearing mutations in genes encoding various parts of the flagellum. Genes involved in nutrient biosynthesis, membrane associated proteins, and a number of conserved hypothetical protein (CHP) genes were also identified. A mutation of one CHP gene caused a positive increase in in planta bacterial growth. This rapid and inexpensive screening method allows the discovery of genes important for in vitro traits that can be correlated to roles in the plant interaction

Multiprobabilistic prediction in early medical diagnoses

This paper describes the methodology of providing multiprobability predictions for proteomic mass spectrometry data. The methodology is based on a newly developed machine learning framework called Venn machines. Is allows to output a valid probability interval. The methodology is designed for mass spectrometry data. For demonstrative purposes, we applied this methodology to MALDI-TOF data sets in order to predict the diagnosis of heart disease and early diagnoses of ovarian cancer and breast cancer. The experiments showed that probability intervals are narrow, that is, the output of the multiprobability predictor is similar to a single probability distribution. In addition, probability intervals produced for heart disease and ovarian cancer data were more accurate than the output of corresponding probability predictor. When Venn machines were forced to make point predictions, the accuracy of such predictions is for the most data better than the accuracy of the underlying algorithm that outputs single probability distribution of a label. Application of this methodology to MALDI-TOF data sets empirically demonstrates the validity. The accuracy of the proposed method on ovarian cancer data rises from 66.7 % 11 months in advance of the moment of diagnosis to up to 90.2 % at the moment of diagnosis. The same approach has been applied to heart disease data without time dependency, although the achieved accuracy was not as high (up to 69.9 %). The methodology allowed us to confirm mass spectrometry peaks previously identified as carrying statistically significant information for discrimination between controls and cases.

Heterogeneous tensor decomposition for clustering via manifold optimization

Tensor clustering is an important tool that exploits intrinsically rich structures in real-world multiarray or Tensor datasets. Often in dealing with those datasets, standard practice is to use subspace clustering that is based on vectorizing multiarray data. However, vectorization of tensorial data does not exploit complete structure information. In this paper, we propose a subspace clustering algorithm without adopting any vectorization process. Our approach is based on a novel heterogeneous Tucker decomposition model taking into account cluster membership information. We propose a new clustering algorithm that alternates between different modes of the proposed heterogeneous tensor model. All but the last mode have closed-form updates. Updating the last mode reduces to optimizing over the multinomial manifold for which we investigate second order Riemannian geometry and propose a trust-region algorithm. Numerical experiments show that our proposed algorithm compete effectively with state-of-the-art clustering algorithms that are based on tensor factorization.

Basic aminopeptidase activity is an emerging biomarker in collagen-induced rheumatoid arthritis

The objective of this study was to investigate the catalytic activity of basic aminopeptidase (APB) and its association with periarticular edema and circulating tumor necrosis factor (TNF)-alpha and type II collagen (CII) antibodies (AACII) in a rat model of rheumatoid arthritis (RA) induced by CII (CIA). Edema does not occur in part of CH-treated, even when AACII is higher than in control. TNF-alpha is detectable only in edematous CII-treated. APB in synovial membrane is predominantly a membrane-bound activity also present in soluble form and with higher activity in edematous than in non-edematous CH-treated or control. Synovial fluid and blood plasma have lower APB in non-edematous than in edematous CII-treated or control. In peripheral blood mononuclear cells (PBMCs) the highest levels of APB are found in soluble form in control and in membrane-bound form in non-edematous CII-treated. CII treatment distinguishes two categories of rats: one with arthritic edema, high AACII, detectable TNF-alpha, high soluble and membrane-bound APB in synovial membrane and low APB in the soluble fraction of PBMCs, and another without edema and with high AACII, undetectable TNF-alpha, low APB in the synovial fluid and blood plasma and high APB in the membrane-bound fraction of PBMCs. Data suggest that APB and CIA are strongly related. (C) 2011 Elsevier B.V. All rights reserved.

Expression Patterns of Cell Adhesion Molecules in Mice`s Lung After Administration of Meso-2,3-Dimercaptosuccinic Acid-Coated Maghemite Nanoparticles

We studied the expression pattern of cell adhesion molecules associated to transendothelial migration of leukocytes in different lung`s vascular compartments after administration of a magnetic fluid sample containing maghemite nanoparticles surface-coated with meso-2,3-dimercaptosuccinic acid. The analyses were conducted in mice 4 and 12 h after endovenous administration of the magnetic fluid in control mice. Firstly, the migratory activity of leukocytes after magnetic fluid surface-coated with meso-2,3-dimercaptosuccinic acid administration was confirmed using broncho-alveolar lavage and light microscopy. Then, the expression of cell adhesion molecules in the lung`s vascular compartments was investigated by immunofluorescence microscopy of frozen sections, using antibodies against L-selectin, P-selectin, E-selectin, macrophage antigen-1, and leukocyte function associated antigen-1. L- and P-selectin showed similar pattern of expression in the pulmonary vasculature in animals treated with magnetic fluid and in the control group. In contrast, macrophage antigen-1 and leukocyte function associated antigen-1 were found in capillary only in animals treated with magnetic fluid surface-coated with meso-2,3-dimercaptosuccinic acid administration. In addition, after magnetic fluid administration E-selectin was found in post-capillary sites. Our findings demonstrated that magnetic fluid surface-coated with meso-2,3-dimercaptosuccinic acid administration exhibits modulation effects on expression patterns of E-selectin, macrophage antigen-1, and leukocyte function associated antigen-1 in the lung`s vascular compartments. These findings are very important in a strategy to reduce the potential toxicity of magnetic fluid surface-coated with meso-2,3-dimercaptosuccinic acid administration for medical applications.

Bradykinin inducible receptor is essential to lipopolysaccharide-induced acute lung injury in mice

Lipopolysaccharides from gram-negative bacteria are amongst the most common causative agents of acute lung injury, which is characterized by an inflammatory response, with cellular infiltration and the release of mediators/cytokines. There is evidence that bradykinin plays a role in lung inflammation in asthma but in other types of lung inflammation its role is less clear. In the present study we evaluated the role of the bradykinin B(1) receptor in acute lung injury caused by lipopolysaccharide inhalation and the mechanisms behind bradykinin actions participating in the inflammatory response. We found that in C57BI/6 mice, the bradykinin B(1) receptor expression was up-regulated 24 h after lipopolysaccharide inhalation. At this time, the number of cells and protein concentration were significantly increased in the bronchoalveolar lavage fluid and the mice developed airway hyperreactivity to methacholine. In addition, there was an increased expression of tumor necrosis factor-alpha, interleukin-1 beta and interferon-gamma and chemokines (monocytes chemotactic protein-1 and KC) in the bronchoalveolar lavage fluid and in the lung tissue. We then treated the mice with a bradykinin B, receptor antagonist, R-954 (Ac-Orn-[Oic(2), alpha-MePhe(5), D-beta Nal(7), Ile(8)]desArg(9)-bradykinin), 30 min after lipopolysaccharide administration. We observed that this treatment prevented the airway hyperreactivity as well as the increased cellular infiltration and protein content in the bronchoalveolar lavage fluid. Moreover, R-954 inhibited the expression of cytokines/chemokines. These results implicate bradykinin, acting through B(1) receptor, in the development of acute lung injury caused by lipopolysaccharide inhalation. (C) 2010 Elsevier B.V. All rights reserved.

PLANT LEAF IDENTIFICATION BASED ON VOLUMETRIC FRACTAL DIMENSION

Texture is an important visual attribute used to describe the pixel organization in an image. As well as it being easily identified by humans, its analysis process demands a high level of sophistication and computer complexity. This paper presents a novel approach for texture analysis, based on analyzing the complexity of the surface generated from a texture, in order to describe and characterize it. The proposed method produces a texture signature which is able to efficiently characterize different texture classes. The paper also illustrates a novel method performance on an experiment using texture images of leaves. Leaf identification is a difficult and complex task due to the nature of plants, which presents a huge pattern variation. The high classification rate yielded shows the potential of the method, improving on traditional texture techniques, such as Gabor filters and Fourier analysis.

Multilevel Training of Binary Morphological Operators

The design of binary morphological operators that are translation-invariant and locally defined by a finite neighborhood window corresponds to the problem of designing Boolean functions. As in any supervised classification problem, morphological operators designed from a training sample also suffer from overfitting. Large neighborhood tends to lead to performance degradation of the designed operator. This work proposes a multilevel design approach to deal with the issue of designing large neighborhood-based operators. The main idea is inspired by stacked generalization (a multilevel classifier design approach) and consists of, at each training level, combining the outcomes of the previous level operators. The final operator is a multilevel operator that ultimately depends on a larger neighborhood than of the individual operators that have been combined. Experimental results show that two-level operators obtained by combining operators designed on subwindows of a large window consistently outperform the single-level operators designed on the full window. They also show that iterating two-level operators is an effective multilevel approach to obtain better results.

Speech Assessment for the Classification of Hypokinetic Dysthria in Parkinson Disease

The aim of this thesis is to investigate computerized voice assessment methods to classify between the normal and Dysarthric speech signals. In this proposed system, computerized assessment methods equipped with signal processing and artificial intelligence techniques have been introduced. The sentences used for the measurement of inter-stress intervals (ISI) were read by each subject. These sentences were computed for comparisons between normal and impaired voice. Band pass filter has been used for the preprocessing of speech samples. Speech segmentation is performed using signal energy and spectral centroid to separate voiced and unvoiced areas in speech signal. Acoustic features are extracted from the LPC model and speech segments from each audio signal to find the anomalies. The speech features which have been assessed for classification are Energy Entropy, Zero crossing rate (ZCR), Spectral-Centroid, Mean Fundamental-Frequency (Meanf0), Jitter (RAP), Jitter (PPQ), and Shimmer (APQ). Naïve Bayes (NB) has been used for speech classification. For speech test-1 and test-2, 72% and 80% accuracies of classification between healthy and impaired speech samples have been achieved respectively using the NB. For speech test-3, 64% correct classification is achieved using the NB. The results direct the possibility of speech impairment classification in PD patients based on the clinical rating scale.

Mechanising first-order temporal resolution

First-order temporal logic is a concise and powerful notation, with many potential applications in both Computer Science and Artificial Intelligence. While the full logic is highly complex, recent work on monodic first-order temporal logics has identified important enumerable and even decidable fragments. Although a complete and correct resolution-style calculus has already been suggested for this specific fragment, this calculus involves constructions too complex to be of practical value. In this paper, we develop a machine-oriented clausal resolution method which features radically simplified proof search. We first define a normal form for monodic formulae and then introduce a novel resolution calculus that can be applied to formulae in this normal form. By careful encoding, parts of the calculus can be implemented using classical first-order resolution and can, thus, be efficiently implemented. We prove correctness and completeness results for the calculus and illustrate it on a comprehensive example. An implementation of the method is briefly discussed.

Monodic Temporal Resolution

First-order temporal logic is a coincise and powerful notation, with many potential applications in both Computer Science and Artificial Intelligence. While the full logic is highly complex, recent work on monodic first-order temporal logics have identified important enumerable and even decidable fragments. In this paper we present the first resolution-based calculus for monodic first-order temporal logic. Although the main focus of the paper is on establishing completeness result, we also consider implementation issues and define a basic loop-search algorithm that may be used to guide the temporal resolution system.