John MacIntyre

A multiobjective evolutionary setting for feature selection and a commonality-based crossover operator

Feature selection is a common and key problem in many classification and regression tasks. It can be viewed as a multiobjective optimisation problem, since, in the simplest case, it involves feature subset size minimisation and performance maximisation. This paper presents a multiobjective evolutionary approach for feature selection. A novel commonality-based crossover operator is introduced and placed in the multiobjective evolutionary setting. This specialised operator helps to preserve building blocks with promising performance. Selection bias reduction is achieved by resampling. We argue that this is a generic approach, which can be used in many modelling problems. It is applied to feature selection on different neural network architectures. Results from experiments with benchmarking data sets are given.

Knowledge extraction from radial basis function networks and multilayer perceptrons

This paper deals with an evaluation and comparison of the accuracy and complexity of symbolic rules extracted from radial basis function networks and multilayer perceptrons. Here we examine the ability of rule extraction algorithms to extract meaningful rules that describe the overall performance of a particular network. In addition, the paper also highlights the suitability of a specific neural network architecture for particular classification problems. The study carried out on the extracted rule quality and complexity also has a direct bearing on the use of rule extraction algorithms for data mining and knowledge discovery.

Flexible software for condition monitoring, incorporating novelty detection and diagnostics

Condition monitoring and machinery fault diagnosis are central to the implementation of efficient maintenance management strategies. They can be based on empirical modelling, which aims at associating measured data to machine conditions. Arguably, different monitoring tasks present different challenges to the maintenance engineer. This paper presents the development of a flexible software solution for condition monitoring, novelty identification and machinery diagnostics, which can easily be customised to a wide range of monitoring scenarios. Its main constituents are a number of independent software modules, such as the fault and symptom tree, the fuzzy classification module, the novelty detection and the neural network diagnostics sub-systems. It is implemented on two different applications, namely machine tool monitoring and gearbox monitoring.

Multiple-criteria genetic algorithms for feature selection in neuro-fuzzy modeling

This paper discusses the use of multicriteria genetic algorithms for feature selection in classification problems. This feature selection approach is shown to yield a diverse population of alternative feature subsets with various accuracy/complexity trade-off. The algorithm is applied to select features for performing classification with fuzzy models, and is evaluated on two real-world data sets. We discuss when multicriteria genetic algorithm feature selection is preferable to a sequential feature selection procedure, namely backwards elimination. Among the key features of the presented approach are its computational simplicity, effectiveness on real world problems and the potential it has to become a powerful tool aiding many empirical modeling and data mining processes.

A Case-Based Reasoning Tool For Vibration Analysis

This paper describes the development of a case-based reasoning (CBR) tool for vibration analysis, the Vibration Case Library (VCL). The system is to help practicing engineers access similar cases while attempting to diagnose actual and potential faults on machines. Of especial interest is the novel calculation of the similarity metric in the complex domain of vibration analysis. This is achieved by means of optimisation of weights based upon analysis of retrieval accuracy (rank ordered lists), using a variant of the Kendal Tau coefficient. Representation is of complex three-dimensional objects and their environments, and is orientated towards high precision retrieval of cases.

'NEURAL-MAINE': intelligent on-line multiple sensor diagnostics for steam turbines in power generation

Neural-Maine is a project under the European EUREKA-Maine (Maintaining Availability in Europe) initiative. The project is in its first of four years and has begun to develop a system for performing plant diagnosis for complex rotating machines such as steam turbines. The key advance involves the use of artificial neural networks for local sensory fusion of multiple transducers reporting to a committee network structure for interpretation and whole plant condition modelling. The work is being carried out by a consortia from the UK and Holland. The paper presents results from the feasibility study and the preliminary work in the development phase.

Applications of neural computing in the twenty-first century and 21 years of Neural Computing & Applications

It is with great pleasure that I am able to write this piece for the twenty-first anniversary edition of Neural Computing & Applications, with our Journal in good health, increasing in stature and reputation year-on-year, and established as a significant publication in the world of computer science and artificial intelligence. As we enter our twenty-second year of publication, I felt this would be a good opportunity to reflect on the changes that have taken place in the tools, techniques, and applications of neural networks and associated adaptive computing technologies, as seen through the submissions to and published papers in our Journal over its 21-year history, and to offer some observations on where we might be going in the future. I hope readers find this reverie interesting, enlightening, or at least entertaining—and are patient enough to indulge me!

Integrating gene expression data from microarrays using the self-organising map and the gene ontology

The self-organizing map (SOM) is useful within bioinformatics research because of its clustering and visualization capabilities. The SOM is a vector quantization method that reduces the dimensionality of original measurement and visualizes individual tumor sample in a SOM component plane. The data is taken from cDNA microarray experiments on Diffuse Large B-Cell Lymphoma (DLBCL) data set of Alizadeh. The objective is to get the SOM to discover biologically meaningful clusters of genes that are active in this particular form of cancer. Despite their powers of visualization, SOMs cannot provide a full explanation of their structure and composition without further detailed analysis. The only method to have gone someway towards filling this gap is the unified distance matrix or U-matrix technique. This method will be used to provide a better understanding of the nature of discovered gene clusters. We enhance the work of previous researchers by integrating the clustering results with the Gene Ontology for deeper analysis of biological meaning, identification of diversity in gene expression of the DLBCL tumors and reflecting the variations in tumor growth rate.

Adaptive Local Fusion Systems for Novelty Detection and Diagnostics in Condition Monitoring

This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine condition monitoring application which learns fault conditions over time. The authors describe the implementation of a novelty detection and adaptive diagnostic system which forms a modular component of a larger on-line condition monitoring system.

Maintenance Strategy Development Within SME’s: The Development of an Integrated Approach

The importance of maintenance has increased, as high productivity and quality can be achieved by means of well-developed and organised maintenance strategies. However, this assumes that maintenance is controlled in such a way that equipment is stopped for maintenance via a systematic schedule. With the recent advances in technology many methodologies, tools, techniques and strategies have been developed and tested. Unfortunately, the majority of organisations are constrained by certain barriers with the resulting loss of major benefits. These are usually classified as Small and Medium Sized Enterprises (SMEs). Based upon our data analysis a new maintenance methodology, the Advanced Integrated Maintenance Management System (AIMMS) is developed. To enable the implementation, monitoring and evaluation of AIMMS a computerised system — Maintenance Management (MainMan) — was developed and implemented within several case study companies. This paper examines the implementation process within one of these companies. The results indicate that AIMMS supports strategic maintenance decisions, and helps to increase equipment effectiveness through prioritising equipment criticality and focusing on specific resources that will maximise gains based upon a return on investment.Copyright