Decision makers are required to make critical decisions throughout all stages of a life-cycle in large-scale projects. These decisions are important as they impact upon the outcome and the success of projects. In this paper we present an evidential reasoning framework to aid decision-makers in the decision making process. This approach utilizes the Dezert-Smarandache Theory (DSm) to fuse heterogeneous evidence sources that suffer from levels of uncertainty, imprecision and conflicts to provide beliefs for decision options. To analyze the impact that source reliability and priority has upon the decision making process a reliability discounting technique along with a priority discounting technique are applied. Application of the evidential reasoning framework is illustrated using a Case Study based in the Aerospace domain.

The seven papers in this special section focus on domain-driven data mining. One paper addresses the general framework of domain-driven data mining, four papers cover the approaches, and two papers consider the workable applications.

Web databases are now pervasive. Such a database can be accessed via its query interface (usually HTML query form) only. Extracting Web query interfaces is a critical step in data integration across multiple Web databases, which creates a formal representation of a query form by extracting a set of query conditions in it. This paper presents a novel approach to extracting Web query interfaces. In this approach, a generic set of query condition rules are created to define query conditions that are semantically equivalent to SQL search conditions. Query condition rules represent the semantic roles that labels and form elements play in query conditions, and how they are hierarchically grouped into constructs of query conditions. To group labels and form elements in a query form, we explore both their structural proximity in the hierarchy of structures in the query form, which is captured by a tree of nested tags in the HTML codes of the form, and their semantic similarity, which is captured by various short texts used in labels, form elements and their properties. We have implemented the proposed approach and our experimental results show that the approach is highly effective.

Developing a desirable framework for handling inconsistencies in software requirements specifications is a challenging problem. It has been widely recognized that the relative priority of requirements can help developers to make some necessary trade-off decisions for resolving con- flicts. However, for most distributed development such as viewpoints-based approaches, different stakeholders may assign different levels of priority to the same shared requirements statement from their own perspectives. The disagreement in the local levels of priority assigned to the same shared requirements statement often puts developers into a dilemma during the inconsistency handling process. The main contribution of this paper is to present a prioritized merging-based framework for handling inconsistency in distributed software requirements specifications. Given a set of distributed inconsistent requirements collections with the local prioritization, we first construct a requirements specification with a prioritization from an overall perspective. We provide two approaches to constructing a requirements specification with the global prioritization, including a merging-based construction and a priority vector-based construction. Following this, we derive proposals for handling inconsistencies from the globally prioritized requirements specification in terms of prioritized merging. Moreover, from the overall perspective, these proposals may be viewed as the most appropriate to modifying the given inconsistent requirements specification in the sense of the ordering relation over all the consistent subsets of the requirements specification. Finally, we consider applying negotiation-based techniques to viewpoints so as to identify an acceptable common proposal from these proposals.

When an agent observes its environment, there are two important characteristics of the perceived information. One is the relevance of information and the other is redundancy. The irrelevant and redundant features which commonly exists within an environment, commonly leads to agent state explosion and associated high computational cost within the learning process. This paper presents an efficient method concerning both the relevance of information and the correlation in order to improve the learning of reinforcement learning agent. We introduce a new concurrent online learning method to calculate the match count C(s) and relevance degree I(s) to quantify the redundancy and correlation of features with respect to a desired learning task. Our analysis shows that the correlation relationship of the features can be extracted and projected to concurrent biased learning threads. By comparing the commonalities of these learning threads, we can evaluate the relevance degree of a feature that contributes to a particular learning task. We explain the method using random walk examples and then demonstrate the method on the chase object domain. Our validation results show that, using the concurrent learning method, we can efficiently detect redundancy and irrelevant features from the environment on sequential tasks, and significantly improve the efficiency of learning. After relevant features are extracted, the agent can remarkably accelerate its succeeding learning speed.

In this paper we describe a system for capturing time-varying sequences of behavioural patterns of agents or robots from a video clips. The detailed time slices of action are 'coarsened' to provide gross, molecular units of behaviour. Combinations of these are then represented in the form of a table, which is mined using rough sets methods here, although various other mining techniques can also be used. Some preliminary experiments are reported to prove the concept.

When symbolic AI approaches are applied to handle continuous valued attributes, there is a requirement to transform the continuous attribute values to symbolic data. In this paper, a novel distribution-index-based discretizer is proposed for such a transformation. Based on definitions of dichotomic entropy and a compound distributional index, a simple criterion is applied to discretize continuous attributes adaptively. The dichotomic entropy indicates the homogeneity degree of the decision value distribution, and is applied to determine the best splitting point. The compound distributional index combines both the homogeneity degrees of attribute value distributions and the decision value distribution, and is applied to determine which interval should be split further; thus, a potentially improved solution of the discretization problem can be found efficiently. Based on multiple reducts in rough set theory, a multiknowledge approach can attain high decision accuracy for information systems with a large number of attributes and missing values. In this paper, our discretizer is combined with the multiknowledge approach to further improve decision accuracy for information systems with continuous attributes. Experimental results on benchmark data sets show that the new discretizer can improve not only the multiknowledge approach, but also the naive Bayes classifier and the C5.0 tree

Knowledge representation is very important in intelligent systems -e.g. for knowledge discovery, data mining, and machine learning. The human brain, a significant intelligent system, works with a huge number of spiking neurons. Based on spiking neuron models a new generation of spiking neural networks (SNNs) has been developed for artificial intelligence systems. SNNs are computationally more powerful than conventional artificial neural networks. In this paper, the spiking neuron model is applied to represent logic rules and fuzzy rules. Based on the STDP (Spike Timing Dependent Plasticity) principle, a new SNN model is proposed for pattern recognition. An efficient learning rule derived from the STDP is applied for self-organizing the input training set efficiently. An example, Animal-Growth-Record, is used to explain the principle of the SNN model. Benchmark data sets are applied to compare the proposed approach with other approaches. As there are very efficient learning rules in the SNN model, the model can be applied not only for fusion of multi-sensory data, but also for data mining in large databases with large numbers of attributes.

A self-organizing computing network based on concepts of fuzzy conditions, beliefs, probabilities, and neural networks is proposed for decision-making in intelligent systems which are required to handle data sets with a diversity of data types. A sense-function with a sense-range and fuzzy edges is defined as a transfer function for connections from the input layer to the hidden layer in the network. By generating hidden cells and adjusting the parameters of the sense-functions, the network self-organizes and adapts to a training set. Computing cells in the input layer are designed as data converters so that the network can deal with both symbolic data and numeric data. Hidden computing cells in the network can be explained via fuzzy rules in a similar manner to those in fuzzy neural networks. The values in the output layer can be explained as a belief distribution over a decision space. The final decision is made by means of the winner-take-all rule. The approach was applied to a series of the benchmark data sets with a diversity of data types and comparative results obtained. Based on these results, the suitability of a range of data types for processing by different intelligent techniques was analyzed, and the results show that the proposed approach is better than other approaches for decision-making in information systems with mixed data types.

In this paper, an evolving learning mechanism is proposed for general computing network model to make decisions in intelligent systems. The novel mechanism is performed by means of a set of computing cell operations such as self-generation, growth, self-division, and death. Under the mechanism, a computing network grows up to a mature network. A hidden cell in the network is defined as a condition matching-unit in response to a fuzzy sub-superspace in multiple-dimension input superspace. A sense-function is defined to represent connections from a hidden cell to input cells. The range and edge vagueness of the sense-function are determined by evolving learning mechanism when sample instances are presented to the network. This network is able to learn from a very few training instances to make decisions for unseen instances. The benchmark data sets from the UCI machine learning repository are applied to test the network and comparable results are obtained.

Experience shows that different text classification methods can give different results. We look here at a way of combining the results of two or more different classification methods using an evidential approach. The specific methods we have been experimenting with in our group include the support vector machine, kNN (nearest neighbors), kNN model-based approach (kNNM), and Rocchio methods, but the analysis and methods apply to any methods. We review these learning methods briefly, and then we describe our method for combining the classifiers. In a previous study, we suggested that the combination could be done using evidential operations and that using only two focal points in the mass functions gives good results. However, there are conditions under which we should choose to use more focal points. We assess some aspects of this choice from an reasoning perspective and suggest a refinement of the approach.

An evidence theoretic classification method is proposed in this paper. In order to classify a pattern we consider its neighbours, which are taken as parts of a single source of evidence to support the class membership of the pattern. A single mass function or basic belief assignment is then derived, and the belief function and the pignistic (‘betting rates’) probability function can be calculated. Then the (posterior) conditional pignistic probability function is calculated and used to decide the class label for the pattern. It is shown that such a classifier extends the standard majority voting based k-nearest neighbour classifier, and it is an approximation to the optimal Bayes classifier. In experiments this classifier performed as well as or better than the voting and distance weighted k-nearest neighbours classifiers with best k, and its performance became stable when the number of neighbours considered was >4.

In transaction processing, an association is said to exist between two sets of items when a transaction containing one set is likely to also contain the other. In information retrieval, an association between two sets of keywords occurs when they cooccur in a document. Similarly, in data mining, an association occurs when one attribute set occurs together with another. As the number of such associations may be large, maximal association rules are sought, e.g., Feldman et al. (1997, 1998). Rough set theory is a successful tool for data mining. By using this theory, rules similar to maximal associations can be found. However, we show that the rough set approach to discovering knowledge is much simpler than the maximal association method.

In this paper we design an experiment which can be depicted as a simple scenario, a very limited ‘world’. In this world, there are an actor that can pursue a project and an observer that is keeping its eyes on the actor. We try to show in the experiment that the observer can to some degree understand the actor based on its knowledge and some metaphors, i.e. understand what the actor is doing and why. As the conclusion of this experiment, we try to show some features of ‘understanding’. These are (1) that ‘understanding’ has to be based on some preliminary knowledge; (2) that ‘understanding’ is a process of incremental learning; (3) that, as for symbolic systems, some metaphors are necessary for mapping real entities into concepts in mind.

Extracting pertinent and useful information from customers has long plagued the process of requirements elicitation. This paper presents a new approach to support the elicitation process. This approach combines various techniques for requirements elicitation which include model-based concept acquisition, goal-driven structured interview and concept reuse. Compared with the available approaches for requirements elicitation, the most significant feature of our approach is that it supports both the automation of interaction with customers by using domain terminology, not software terminology and the automated construction of application requirements models using model-based concept elicitation and concept reuse. The capacity of this approach comes from its rich knowledge which is clustered into several abstract levels.

This paper presents a novel method for classification, which makes use of models built by the lattice machine (LM). The LM approximates data resulting in, as a model of data, a set of hyper tuples that are equilabelled, supported and maximal. The method presented uses the LM model of data to classify new data with a view to maximising the density of the model. Experiments show that this method, when used with the LM, outperforms the C2 algorithm and is comparable to the C5.0 classification algorithm.

This paper compares three different data analysis methods in a subfield of the coronary heart disease risk assessment (CHDRA) area—the identification of increased blood cholesterol levels. A data set containing the cholesterol data of 166 persons is employed as a test case, and analyzed in three experimental investigations. The first analysis method employs a predominantly knowledge-based fuzzy expert system solution to the problem. The second method employs statistical discriminant analysis on the same data, whereas the results of the third analysis are obtained by an artificial neural network. This study evaluates the advantages as well as the disadvantages of each method. Special attention is given to the systems' individual capability for managing the uncertainty involved in the decision-making process. The results achieved in the study provide evidence for the complementary and mutually supporting character of the different approaches. ©2000 John Wiley & Sons, Inc.

Within the framework of multilayer feedforward networks, detailed design of the network-number of layers and the number of units within each layer-has long been a concern. The trial and error process of adding units and/or layers to the network is successful in sofar as it often achieves near-optimal performance but it is very time consuming. Constructive algorithms would be the ideal solution if the resulting networks generalised as well as their trial-and-error counterparts. We examine the tiling algorithm of Mezard and Nadal (1989) in greater detail. By choosing various alternative strategies for training the ancillary units in each layer, we are able to reduce the size of the network and improve generalisation.

Relevance has traditionally been linked with feature subset selection, but formalization of this link has not been attempted. In this paper, we propose two axioms for feature subset selection-sufficiency axiom and necessity axiom-based on which this link is formalized: The expected feature subset is the one which maximizes relevance. Finding the expected feature subset turns out to be NP-hard. We then devise a heuristic algorithm to find the expected subset which has a polynomial time complexity. The experimental results show that the algorithm finds good enough subset of features which, when presented to C4.5, results in better prediction accuracy.

Retrieving huge collections of images from the World Wide Web (WWW) based on image content is a challenging field which is receiving a lot of attention from researchers. Most investigations focus on feature extraction and feature matching motivated by image processing and computer vision. We propose a distributed content-based image query system (DCBIQ) based on the WWW. The system architecture integrating knowledge from image processing, semantic descriptors, multi-agents, and WWW navigation is proposed. By extracting the texture, colour, shape features and spatial relationships of the image, we use semantic descriptor models, along with constraints, to query desired images distributed on the WWW. An initial prototype based on NT and extended SQL server has been implemented.

This paper reports on the design, development and evaluation of a buffer algorithm named RESBAL, which exploits parallelism in order to provide high query execution performance in relational database systems. The algorithm aims to provide both predictive and efficient data buffering by exploiting the use of advance knowledge of query reference behaviour. Designed to offer a high level of flexibility, RESBAL employs a multiple buffering strategy both on page fetch level and page replacement level in order to improve buffer performance. The evaluation of RESBAL has been carried out in a parallel database system environment based on a transputer architecture. The results of this performance assessment allow comparison to be made between different buffer algorithms, and demonstrate the feasibility and effectiveness of the RESBAL algorithm. © 1998 John Wiley & Sons, Ltd.

Rough set theory is a new mathematical tool to deal with vagueness and uncertainty. To apply the theory, it is important to associate it with efficient and effective computational methods. With a little adjustment, a relation can be used to represent a decision table for use in decision making. By using this kind of table, rough set theory can be applied successfully to rough classification and knowledge discovery. We present computational methods for using rough sets to identify classes in datasets, finding dependencies in relations, and discovering rules which are hidden in databases. The methods are illustrated with a running example from a database of car test results. © 1998 John Wiley & Sons, Inc.

This paper presents an efficient algorithm for learning Bayesian belief networks from databases.The algorithm takes a database as input and constructs the belief network structure as output.The construction process is based on the computation of mutual information of attribute pairs.Given a data set that is large enough, this algorithm can generate a belief network very close to the underlying model, and at the same time, enjoys the time complexity of 0(N4) on conditional independence (CI) tests.When the data set has a normal DAG-Faithfil (see Section 3.2) probability distribution, the algorithm guarantees that the structure of a perfect map pearl, 19881 of the underlying dependency mode1 is generated.To evaluate this algorithm, we present the experimental results on three versions of the wellknown ALARM nehvork database, which has 37 attributes and 10,000 records.The results show that this algorithm is accurate and efficient.The proof of correctness and the analysis of computational complexity are also presented. 1 -*. 1 .

Describes a practicable approach to extend relational databases to deductive databases by using a two-level knowledge representation mechanism. By this approach, we developed a deductive database add-in module that can be easily integrated with relational databases. Therefore, users can get deductive reasoning functionality without changing their favourite relational databases and query languages. This approach proved to be efficient on knowledge base organization and reasoning in real-world applications.

The ideal situation for a Data Mining or Knowledge Discovery system would be for the user to be able to pose a query in the actuarial domain.

Abstract There is a perceived need within the database community to extend the traditional relational database systems so as to accommodate applications which are deductive in nature. One major problem involved in such an extension is the efficient processing of recursive queries. To this end, parallel processing is expected to play an important role. While substantial work has been done in devising strategies for processing recursive queries in parallel, it is perhaps surprising that little has been reported on the implementation and the run‐time performance of these strategies. In the paper we report our experience of implementing a pipelined evaluation strategy on transputers. A wide range of queries, database structures and architectural configurations are considered as benchmarks in this study. The performance is studied in terms of both speed‐up factors and communication costs. The experimental results show the potential of processing recursive queries in parallel, and provide insight into the usefulness of using transputers for such applications.

The NavChair assistive navigation system is being developed to meet the needs of multiply handicapped people who are unable to operate available wheelchair systems. The NavChair Project was conceived as an application of mobile robot obstacle avoidance to a power wheelchair. During the course of this project, the vector field histogram (VFH) method has been adapted for use in human-machine systems and the shortcomings of the wheelchair platform have been overcome. This paper briefly reviews the VFH method, describes interesting aspects of its application to a power wheelchair, and presents an experimental evaluation of system performance. Finally, unresolved problems of obstacle avoidance in human-operated vehicles are discussed.>

This paper presents a framework for fault diagnosis in distributed expert systems. We investigate computational methods for evidential reasoning in expert systems based on the Guan-Bell approach. Matrices and their products are used to express the reasoning algorithms. This makes the algorithms simpler and perhaps more efficient to implement. To make the description of our method clear, we trace through the analysis of the simplified problem of an example in the context of fault diagnosis in a distributed vehicle monitoring system.>

The problem of making decisions among propositions based on both uncertain data items and arguments which are not certain is addressed. The primary knowledge discovery issue addressed is a classification problem: which classification does the available evidence support? The method investigated seeks to exploit information available from conventional database systems, namely, the integrity assertions or data dependency information contained in the database. This information allows ranking arguments in terms of their strengths. As a step in the process of discovering classification knowledge, using a database as a secondary knowledge discovery exercise, latent knowledge pertinent to arguments of relevance to the purpose at hand is explicated. This is called evidence. Information is requested via user prompts from an evidential reasoner. It is fed as evidence to the reasoner. An object-oriented structure for managing evidence is used to model the conclusion space and to reflect the evidence structure. The implementation of the evidence structure and an example of its use are outlined.>

This paper describes a new architecture for a system to handle temporal data, called Temporal Data Management System, TDMS. TDMS represents an approach to the modelling of temporal semantics which is pragmatic because it takes into account practical issues such as the time/space trade-offs. We consider the technical aspects of the system i.e. how time domain data is managed in TDMS. The design of temporal databases is not addressed here. We introduce a conceptual framework which permits issues such as handling schema anomalies due to updates, query validation and correct semantics modelling to be handled in a consistent and usable manner. It incorporates the time values of each of three ‘data constructs’ organised in a hierarchy. This is combined at the implementation level with a well-established multi-database architecture to give a comprehensive and efficient temporal model which overcomes several weaknesses of current temporal models. Modest extensions to the SQL syntax to handle time are proposed and have been implemented. Both tuple-stamping and value-stamping are provided by the integration of TDMS and the relational data model. TDMS has been prototyped under the UNIX operating system using the C programming language. Presently, TDMS interfaces with the Ingres DBMS.

One approach to deductive database system implementation which has received much recent attention is the integration of a relational database management system and a logical inference unit. This makes the deductive manipulation of a large amount of data computationally possible. While several proposals for such systems have been made, differing terminologies and implementations make them difficult to compare. This paper presents a systematic study on the integrated approach to deductive database system implementation. In addition to serving as a foundation for the implementation of new systems, this study clarifies and emphasises the necessary features of integrated deductive database systems.

Techniques for general purpose optimization have been derived from the Metropolis Monte Carlo method of simulating the behavior of particles in substances as they are slowly cooled to form crystals. Simulated Annealing is such a derivative and its value for placement problems (e.g., in circuit board layout design) suggests that it could be advantageously applied to clustering tuples in databases in order to enhance responsiveness to queries. In this article we investigate this issue and compare the performance of this technique with a Graph-Collapsing clustering method which is known to perform very well, in order to gain insights into which approach is better for incorporation in a performance-oriented database design tool. We judge that, whilst the new method does give superior results to the graph-based method in many cases, these improvements are gained at such a very considerable expense of algorithm run time as to rule the new technique out of our consideration as a real world general purpose design tool (but perhaps not for some special-purpose databases). © 1990 John Wiley & Sons, Inc.

The causes of deadlocks in distributed expert systems are examined and classified, and two methods of processing them are proposed. In particular, this problem is discussed within the framework of a working prototype distributed expert system, Hecodes (heterogeneous cooperative distributed expert system) which consists of a control system, a blackboard system, and a group of expert systems whose structures and functions can be different. The two approaches for processing any deadlock are termed deadlock avoidance, and deadlock detection and destruction. Only one of the two can be chosen for a particular run of the system. The detailed working and effects of using each of the proposed deadlock resolution methods are discussed.>

A method is presented for modeling attribute value distributions in database relations for the purpose of obtaining accurate estimates of intermediate relation sizes during query evaluation. The basic idea is that instead of keeping a single (average) value to represent the number of occurrences of each attribute value, m (typically ten) parameters are kept, each representing the number of occurrences of attribute values in a piece, or partition, corresponding to a subrange of 1/mth of the original value range. The uniformity assumption, taken as an estimation technique rather than as an assumption, holds for each partition, hence the name piecewise uniform. The distribution method is extended to the modeling of important intrarelational attribute correlations. This and other enhancements to the technique such as application to semijoin operation are suggested. The technique is being used on two multidatabase management systems.>

This paper presents empirical results obtained using a heuristic graph-collapsing technique for placing data on storage devices. Previous experiments using this technique worked on small datasets. We have set out to gain further experience with the graph-collapsing algorithm using more realistic data sizes. These experiments have given us promising results and stimulated ideas which should be beneficial to the database designer in the future.

The problems of data placement are investigated abstractly, and we use the theory of NP-completeness to prove some results which indicate the level of difficulty of some increasingly realistic data placement problems. As a consequence of these theorems it is established that the likelihood of finding a deterministic algorithm to solve any of these data placement problems with acceptable performance is very small, and other, possibly heuristic, methods are suggested. A particularly interesting data placement problem considered in this study is formulated as the layered data placement problem, which is shown to be NP-complete.

The hash trees method of external hashing is known to have advantages for certain types of primary key distribution. In this paper the value of the method as a general indexing technique—for secondary keys as well as primary keys—is assessed, and a comparison with the B-trees method is presented.

The high cost of training programmers prohibits a trial and error approach to selection. This paper investigates the qualities required in a modern programmer and a means of evaluating them. The results of a recent experiment to determine which of a collection of programmer selection tests and subtests proved to be the best predictor are described.

Information Theory for Systems Engineers L. P. Hyvärinen , 1968 ; 205. (No. 5 in Lecture Notes in Operations Research and Mathematical Economics.) (Berlin : Springer-Verlag , $3.80 .)

Error Correcting Codes Henry B. Mann , 1969, 228. (John Wiley & Sons Ltd. , Price 75s .)

Concatenated Codes G. David Forney Jr. , 1966 ; 147 + ix. Research Monograph No. 37, (Massachusetts : M.I.T. Press , 32s.)

This paper discusses the possibility of obtaining the Fourier transform of an arbitrary function, which is given in numerical or graphical form, by approximating it piece-wise with simple analytic functions. The Fourier transform is a necessary tool for translation between the time-function and frequency-function des-cription of a waveform, or between linear distribution of intensity across an aperture and angular distribution of radiation from the aperture. If the function to be transformed is of reasonably simple analytical form, there is no difficulty: one has merely to consult a table of Fourier transforms such as was produced by Campbell and Foster (1948). But if the subject to be transformed is a set of experimental data, not conforming precisely to any analytical law, the problem is less easy. The direct approach is to perform numerically "the integration

A study of one or two specific problems in the application of electronic computers to clerical work has convinced the writer that sorting is one of the clerical processes for which electronic techniques are not yet adequately developed. The following notes seek first to analyse the process of sorting into general principles, which may be useful in the search for further practical solutions of the problem, and then to classify existing sorting techniques through their relationship to the general principles.