Ripple down rules (RDR), that is rules with hierarchical exceptions, are used in knowledge acquisition because they provide a well intelligible and modifiable representation for even very large expert systems. In this paper a formal semantics for RDRs is proposed, that covers first order rules as well as attribute-value based rules. An algebraic foundation is proposed, including simplification of RDRs and transformation of RDRs into flat lists of rules and ripple down rule sets, hence these knowledge representation schemes are put into perspective. It is shown, that a RDR has a shorter description length than an equivalent decision list. Induction of rules with exceptions is characterized as bidirectional movement in the hypothesis space, while known algorithms for learning rules or decision trees either perform a top-down specialization of the most general or a bottom-up generalization of the most special hypothesis. Known algorithms for induction of RDRs are summarized and compared a...

. Knowledge-based systems (KBS) are not necessarily based on well-defined ontologies. In particular it is possible to build KBS for classification problems, where there is little constraint on how classes are organised and a class is expressed by the expert as a free text conclusion to a rule. This paper investigates how relations between such 'classes' may be discovered from existing knowledge bases, then investigates how to construct a model of these classes (an ontology) based on user-selected patterns in the class relations. We have applied our approach to KBS built with Ripple Down Rules (RDR) [1] RDR is a knowledge acquisition and knowledge maintenance methodology, which allows KBS to be built very rapidly and simply, but does not require a strong ontology. Our experimental results are based on a large real-world medical RDR KBS. The motivation for our work is to allow an ontology in a KBS to 'emerge' during development, rather than requiring the ontology to be established prior to the development of the KBS.

The work reported in this thesis is motivated by the belief that knowledge-based systems (KBS) research needs to focus more on users ’ needs and cater for the various decision situations in which users will find themselves. To build individual systems that cater for all the activities that may be needed is not feasible or desirable. The problems associated with capturing knowledge are well known and the ability to capture knowledge once and access and manipulate the knowledge in multiple ways is highly desirable. It adds value to the original knowledge and offers all the benefits associated with the reuse of resources. Thus, the problem becomes one of knowledge reuse. The research question pursued in this thesis is “can knowledge captured for one purpose, such as consultation, be reused to support a wide range of alternative purposes, such as critquing or tutoring, allowing the user to answer different types of questions according to their current circumstances”? Further, this question was to be answered in a situated cognition, dynamic knowledge framework. The system developed in this thesis is based on the Multiple Classification Ripple Down Rule (MCRDR) knowledge acquisition and representation technique. MCRDR is a form of casedbased

Abstract: Ripple Down Rules (RDR) have a longstanding (and successful) history in the field of biomedical engineering. RDR are a knowledge acquisition and representation technique that allow knowledge to be rapidly acquired and maintained by the domain expert. A key feature of RDR, and the reason why maintenance is easily managed, is that rules are never modified or deleted but they are locally patched. That is, new rules are exceptions to previous rules and the new rule is validated within the context of previously seen cases. One drawback of locally patching is that knowledge can be repeated in different locations of the knowledge base. This paper describes some work done on removing repeated knowledge. The experiments reported were performed on a pathology knowledge base but the algorithm is applicable to any multiple classification RDR knowledge based system. The results support the findings of others that exception structures are compact representations with few opportunities to reduce further. This also suggests that experts tend to provide overly general rules in the first instance which they modify by adding specialistions in the form of exception rules as new cases are seen. 1

The search for patterns in data sets is a fundamental task in Data Mining, where Machine Learning algorithms are generally used. However, Machine Learning algorithms have biases that strengthen the classification task, not taking into consideration exceptions. Exceptions contradict common sense rules. They are generally unknown, unexpected and contradictory to the user believes. For this reason, exceptions may be interesting. In this work we propose a method to find exceptions out from common sense rules. Besides, we apply the proposed method in a real world data set, to discover rules and exceptions in the HIV virus protein cleavage process.

Abstract. In this paper we investigate how Ripple Down Rules (RDR) knowledge-based systems (KBS) can be reorganized and intermediate concepts discovered. The experiment is based on a real-world multiple-classification RDR knowledge base (KB) with 3710 rules and 2211 cornerstone cases used for interpreting lipid results in chemical pathology. RDR knowledge acquisition can start with a minimal ontology and adding a rule is simple and rapid. On the other hand there is the drawback of possible repetition, redundancy and lack of intermediate concepts. This paper demonstrates how repetition and redundancy can be elimitated by reorganising RDR and how intermediate concepts can be discovered. 1

Abstract. The aim of this study was to develop machine learning techniques that would speed up knowledge acquisition from an expert. As the expert provided knowledge the system would generalize from this knowledge in order to reduce the need for later knowledge acquisition. This generalization should be completely hidden from the expert. We have developed such a learning technique based on Duce’s intra-construction and absorption operators [1] and applied to Ripple Down Rule (RDR) incremental knowledge acquisition [2]. Preliminary evaluation shows that knowledge acquisition can be reduced by up to 50%. 1

Abstract—Researchers in machine learning use decision trees, production rules, and decision graphs for visualizing classification data. Part I of this paper surveyed these representations, paying particular attention to their comprehensibility for non-specialist users. Part II turns attention to knowledge visualization—the graphic form in which a structure is portrayed and its strong influence on comprehensibility. We analyze the questions that, in our experience, end users of machine learning tend to ask of the structures inferred from their empirical data. By mapping these questions onto visualization tasks, we have created new graphical representations that show the flow of examples through a decision structure. These knowledge visualization techniques are particularly appropriate in helping to answer the questions that users typically ask, and we describe their use in discovering new properties of a data set. In the case of decision trees, an automated software tool has been developed to construct the visualizations. Decision trees, production rules and decision graphs are widely used for representing the results of machine learning. As the first part of this paper showed, many schemes have been developed to assist comprehension by reducing the amount of gratuitous information in such