Exploratory analysis of excitation-emission matrix fluorescence spectra with self-organizing maps—A tutorial

Large datasets are common in chemical and environmental engineering applications and tools for their analysis are in great demand. Here, the outputs of a series of fluorescence spectroscopy analyses are utilised to demonstrate the application of the self-organising map (SOM) technique for data analysis. Fluorescence spectroscopy is a well-established technique of organic matter fingerprinting in water. The technique can provide detailed information on the physico-chemical properties of water. However, analysis of fluorescence spectra requires the application of robust statistical and computational data pre-processing and analysis tools.This paper presents a tutorial for training engineering postgraduate researchers in the use of SOM techniques using MATLAB^®. Via a tutorial, the application of SOM to fluorescence spectra and, in particular, the characterisation of organic matter removal in water treatment, is presented. The tutorial presents a step-by-step example of the application of SOM to fluorescence data analysis and includes the source code for MATLAB^®, together with presentation and discussion of the results. With this tutorial we hope to popularise this robust pattern recognition technique for fluorescence data analysis and large data sets in general, and also to provide educational practitioners with a novel tool with which to train engineering students in SOM.     

On model design for simulation of collective intelligence

The study of collective intelligence (CI) systems is increasingly gaining interest in a variety of research and application domains. Those domains range from existing research areas such as computer networks and collective robotics to upcoming areas of agent-based and insect-based computing; also including applications on the internet and in games and movies. CI systems are complex by nature and (1) are effectively adaptive in uncertain and unknown environments, (2) can organise themselves autonomously, and (3) exhibit ‘emergent’ behaviour. Among others, multi-agent systems, complex adaptive systems, swarm intelligence and self-organising systems are considered to be such systems. The explosive wild growth of research studies of CI systems has not yet led to a systematic approach for model design of these kinds of systems. Although there have been recent efforts on the issue of system design (the complete design trajectory from identifying system requirements up to implementation), the problem of choosing and specifying a good model of a CI system is often done implicitly and sometimes even completely ignored. The aim of this article is to bring to the attention that model design is an essential as well as an integral part of system design. We present a constructive approach to systematically design, build and test models of CI systems. Because simulation is often used as a way to research CI systems, we particularly focus on models that can be used for simulation. Additionally, we show that it is not necessary to re-invent the wheel: here, we show how existing models and algorithms can be used for CI model design. The approach is illustrated by means of two example studies on a (semi-automated) multi-player game and collaborative robotics.     

Detection of Bone Tumours in Radiographic Images using Neural Networks

We develop an approach for segmenting radiographic images of focal bone lesions possibly caused by bone tumour. A neural network is used to classify individual pixels by a convolution operation based on a feature vector. We design eight features that characterise the local texture in the neighbourhood of a pixel. Four of the features are based on co-occurrence matrices computed from the neighbourhood. The true class label of the pixels in the radiographs are obtained from annotations made by an experienced radiologist. Neural networks and self-organising feature maps are trained to perform the segmentation task. The experiments confirm the feasibility of using a feature-based neural network for finding pathologic bone changes in radiographic images. An analysis of the eight features indicates that the presence of edges and transitions, the complexity of the texture, as well as the amount of high frequencies in the texture, are the main features discriminating (soft) tissue from pathologic bone, the two classes most likely to be confused. Receiveed: 4 June 1998?,Received in revised form: 15 September 1999?Accepted: 11 December 1998     

Perceptual primitives from an extended 4D Hough transform

Directional features extracted from Gabor wavelets responses were used to train a structure of self-organising maps, thus classifying each pixel in the image within a neuron-map. Resulting directional primitives were grouped into perceptual primitives introducing an extended 4D Hough transform to group pixels with similar directional features. These can then be used as perceptual primitives to detect salient structures. The proposed method has independently fixed parameters that do not need to be tuned for different kind or quality of images. We present results in application to noisy FLIR images and show that line primitives for complex structures, such as bridges, or simple structures, such as runways, can be found by this approach. We compare and demonstrate the quality of our results with those obtained through a parameter-dependent traditional Canny edge detector and Hough line finding process.     

New learning rules for the ASSOM network

The ASSOM is a self-organising neural network with the capability of adapting to linear subspaces. Here we propose two new methods to train the ASSOM network. A nonlinear system of equations is derived for network training. This system can be solved by a gradient-based approach or by the Levenberg–Marquardt method. Each of these two approaches gives a different learning rule. A comparison is carried out among the original Kohonens method and the proposed learning rules. Experimental results are reported, including a convergence speed experiment and a speech processing application, which show that the new learning rules have better performance than the original one.     

A self-organising mixture autoregressive network for FX time series modelling and prediction

Nowadays a great deal of effort has been made in order to gain advantages in foreign exchange (FX) rates predictions. However, most existing techniques seldom excel the simple random walk model in practical applications. This paper describes a self-organising network formed on the basis of a mixture of adaptive autoregressive models. The proposed network, termed self-organising mixture autoregressive (SOMAR) model, can be used to describe and model nonstationary, nonlinear time series by means of a number of underlying local regressive models. An autocorrelation coefficient-based measure is proposed as the similarity measure for assigning input samples to the underlying local models. Experiments on both benchmark time series and several FX rates have been conducted. The results show that the proposed method consistently outperforms other local time series modelling techniques on a range of performance measures including the mean-square-error, correct trend predication percentage, accumulated profit and model variance.     

Standardising vulnerability categories

Each vulnerability scanner (VS) represents, identifies and classifies vulnerabilities in its own way, thus making the different scanners difficult to study and compare. Despite numerous efforts by researchers and organisations to solve the disparity in vulnerability names used in the different VSs, vulnerability categories have still not been standardised. This paper highlights the importance of having a standard vulnerability category set. It also outlines an approach towards achieving this goal by generating a standard set of vulnerability categories. A data-clustering algorithm that employs artificial intelligence is used for this purpose. The significance of this research results from having an intelligent technique that aids in the generation of standardised vulnerability categories in a relatively fast way. In addition, the technique is generic in the sense that it allows one to accommodate any VS currently known on the market to create such vulnerability categories. Another benefit is that the approach followed in this paper allows one to also compare various VSs currently available on the market. A prototype is presented to verify the concept.     

Externally Growing Cell Structures for Data Evaluation of Chemical Gas Sensors

Based on Fritzke’s GCS (Growing Cell Structures), we present here a new incremental self-organising neural network, the Externally Growing Cell Structures (EGCS). Our goals are to speed up the convergence and to improve the generalisation performance. The mechanism of internally growing cells in EGCS is the same as in GCS. However, when the Maximum Resource Vertex (MRV) or the Maximum Error Vertex (MEV) is a boundary node, the new cell is grown externally. Simulation results on neural network benchmarks, two-spiral problem and sonar mine/rock separation, indicate that EGCS performs better than the original GCS, measured by classification rate and the required number of epochs. As a new classification and regression method, the EGCS for Data Evaluation of Chemical Gas Sensors is introduced.     

A Self-Organising Fuzzy Logic Controller for a Coordinate Machine

For a 3D coordinate measurement system, the dynamic accuracy of the moving table will influence the measuring accuracy directly. If a classical PID controller were designed for this measuring table without an accurate mathematical model, the gain parameters may need to be regulated frequently by trial-and-error to obtain the precise motion control objective, good adaptability, and robustness. In this paper, a model-free fuzzy controller and a self-organising fuzzy controller (SOFC) were employed to eliminate the above controller design problems and improve the tracking control accuracy. The control performances of these intelligent control strategies were compared, based on the experimental results. The SOFC has the best tracking accuracy and its learning ability significantly reduces the trial-and-error design effort of a traditional fuzzy controller.     

Filtering multilingual Web content using fuzzy logic and self-organizing maps

Effective multilingual information filtering is required to alleviate users burden of information overload resulting from the increasing flood of multilingual textual content available extensively over the World-Wide Web. This paper proposes a content-based self-organizing approach to multilingual information filtering using fuzzy logic and the self-organizing map. This approach screens and evaluates multilingual documents based on their semantic contents. Correlated multilingual documents are disseminated according to their corresponding themes or topics, thus enabling language-independent content-based information access efficiently and effectively. A Web-based multilingual online news-filtering system is developed to illustrate how the approach works.