This paper introduces ICET, a new algorithm for cost-sensitive classification. ICET uses a genetic algorithm to evolve a population of biases for a decision tree induction algorithm. The fitness

A surprisingly large number of research disciplines have contributed towards the development of knowledge on lazy problem solving, which is characterized by its storage of ground cases and its demand driven response to queries. Case-based reasoning (CBR) is an alternative, increasingly popular approach for designing expert systems that implements this approach. This paper lists pointers to some contributions in some related disciplines that offer insights for CBR research. We then outline a small number of Navy applications based on this approach that demonstrate its breadth of applicability. Finally, we list a few successful and failed attempts to apply CBR, and list some predictions on the future roles of CBR in applications.

One problem with using CBR for diagnosis is that a full case description may not be available at the beginning of the diagnosis. The standard CBR methodology requires a detailed case description in order to perform case retrieval and this is often not practical in diagnosis. We describe two fault diagnosis tasks where many features may make up a case description but only a few features are required in an individual diagnosis. We evaluate an incremental CBR mechanism that can initiate case retrieval with a skeletal case description and will elicit extra discriminating information during the diagnostic process. Keywords: Case-based reasoning, case retrieval, electronic fault diagnosis. 2 1 Introduction The fact that human problem solving competence is often based on reasoning from examples supports the use of case-based reasoning (CBR) for developing knowledge-based systems. In particular, good performance in both technical and medical diagnosis is often dependent on remembering simi...

The Inductive Logic Programming community has considered proof-complexity and model-complexity, but, until recently, size-complexity has received little attention. Recently a challenge was issued "to the international computing community" to discover low size-complexity Prolog programs for classifying trains. The challenge was based on a problem first proposed by Ryszard Michalski, 20 years ago. We interpreted the challenge as a problem in cost-sensitive classification and we applied a recently developed cost-sensitive classifier to the competition. Our algorithm was relatively successful (we won a prize). This paper presents our algorithm and analyzes the results of the competition.

. This chapter 1 investigates the relation between the two key aspects of intelligent agents: intelligence and agent-hood. Efforts for building general architectures for intelligent agents have typically focussed on one of these aspects. The chapter first reviews some work on cognitive architectures for intelligence and illustrates how these are being applied to physical agents. Secondly, this chapter puts forward a series of hypotheses on the relationship between cognition and agent-hood. These hypotheses are aimed at a unified treatment of behavior and cognition. Central to this discussion in the notion of coordination. In particular the nature and role of representations, internal and external, are explained in relation to their coordinating role. 1 Introduction This chapter is a tour along a number of architectural options for building intelligent agents. The concept of agent refers to a system that can be differentiated from its environment and is capable of direct and continu...

The field of knowledge engineering has been one of the most visible successes of artificial intelligence (AI) to date. Knowledge acquisition (KA) is the main bottleneck in the knowledge engineer's (KE) work. Machine learning (ML) tools have contributed positively to the process of trying to eliminate or open up this bottleneck. But how do we know whether the field is progressing? How can we establish the progress made in any of its branches? How can we be sure of an advance and take advantage of it? This paper proposes a benchmark as a classifying, comparative and metric criterion for ML tools. The benchmark has been centred on the KE viewpoint, covering some of the characteristics he wants to find in a ML tool. The proposed model has been applied to a set of ML tools, comparing expected and obtained results. Experimentation has validated the model and led to interesting results. Keywords: concept learning, benchmark, measures 2 APPLYING METRICS TO ML TOOLS: A KE APPROACH 1.- INTROD...

One problem with using CBR for diagnosis is that a full case description may not be available at the beginning of the diagnosis. The standard CBR methodology requires a detailed case description in order to perform case retrieval and this is often not practical in diagnosis. We describe two fault diagnosis tasks where many features may make up a case description but only a few features are required in an individual diagnosis. We evaluate an incremental CBR mechanism that can initiate case retrieval with a skeletal case description and will elicit extra discriminating information during the diagnostic process. Keywords: Case-based reasoning, case retrieval, electronic fault diagnosis. 2 1 Introduction The fact that human problem solving competence is often based on reasoning from examples supports the use of case-based reasoning (CBR) for developing knowledge-based systems. In particular, good performance in both technical and medical diagnosis is often dependent on remembering simi...

The classification of new cases using a predictive model incurs two types of costs—testing costs and misclassification costs. Recent research efforts have resulted in several novel algorithms that attempt to produce learners that simultaneously minimize both types. In many real life scenarios, however, we cannot afford to conduct all the tests required by the predictive model. For example, a medical center might have a fixed predetermined budget for diagnosing each patient. For cost bounded classification, decision trees are considered attractive as they measure only the tests along a single path. In this work we present an anytime framework for producing decision-tree based classifiers that can make accurate decisions within a strict bound on testing costs. These bounds can be known to the learner, known to the classifier but not to the learner, or not predetermined. Extensive experiments with a variety of datasets show that our proposed framework produces trees with lower misclassification costs along a wide range of testing cost bounds.

recognition genetic algorithm

Abstract not found