Marko Bohanec

Trading Accuracy for Simplicity in Decision Trees

When communicating concepts, it is often convenient or even necessary to define a concept approximately. A simple, although only approximately accurate concept definition may be more useful than a completely accurate definition which involves a lot of detail. This paper addresses the problem: given a completely accurate, but complex, definition of a concept, simplify the definition, possibly at the expense of accuracy, so that the simplified definition still corresponds to the concept “sufficiently” well. Concepts are represented by decision trees, and the method of simplification is tree pruning. Given a decision tree that accurately specifies a concept, the problem is to find a smallest pruned tree that still represents the concept within some specified accuracy. A pruning algorithm is presented that finds an optimal solution by generating a dense sequence of pruned trees, decreasing in size, such that each tree has the highest accuracy among all the possible pruned trees of the same size. An efficient implementation of the algorithm, based on dynamic programming, is presented and empirically compared with three progressive pruning algorithms using both artificial and real-world data. An interesting empirical finding is that the real-world data generally allow significantly greater simplification at equal loss of accuracy.

A function-decomposition method for development of hierarchical multi-attribute decision models

Function decomposition is a recent machine learning method that develops a hierarchical structure from class-labeled data by discovering new aggregate attributes and their descriptions. Each new aggregate attribute is described by an example set whose complexity is lower than the complexity of the initial set. We show that function decomposition can be used to develop a hierarchical multi-attribute decision model from a given unstructured set of decision examples. The method implemented in a system called HINT is experimentally evaluated on a real-world housing loans allocation problem and on the rediscovery of three hierarchical decision models. The experimentation demonstrates that the decomposition can discover meaningful and transparent decision models of high classification accuracy. We specifically study the effects of human interaction through either assistance or provision of background knowledge for function decomposition, and show that this has a positive effect on both the comprehensibility and classification accuracy.

Learning by discovering concept hierarchies

Abstract We present a new machine learning method that, given a set of training examples, induces a definition of the target concept in terms of a hierarchy of intermediate concepts and their definitions. This effectively decomposes the problem into smaller, less complex problems. The method is inspired by the Boolean function decomposition approach to the design of switching circuits. To cope with high time complexity of finding an optimal decomposition, we propose a suboptimal heuristic algorithm. The method, implemented in program HINT (Hierarchy INduction Tool), is experimentally evaluated using a set of artificial and real-world learning problems. In particular, the evaluation addresses the generalization property of decomposition and its capability to discover meaningful hierarchies. The experiments show that HINT performs well in both respects.

Data mining and visualization for decision support and modeling of public health-care resources

This paper proposes an innovative use of data mining and visualization techniques for decision support in planning and regional-level management of Slovenian public health-care. Data mining and statistical techniques were used to analyze databases collected by a regional Public Heath Institute. We also studied organizational aspects of public health resources in the selected Celje region with the objective to identify the areas that are atypical in terms of availability and accessibility of public health services for the population. The most important step was the detection of outliers and the analysis of availability and accessibility deviations. The results are applicable to health-care planning and support in decision making by local and regional health-care authorities. In addition to the practical results, which are directly useful for decision making in planning of the regional health-care system, the main methodological contribution of the paper are the developed visualization methods that can be used to facilitate knowledge management and decision making processes.

Knowledge-based portfolio analysis for project evaluation

Abstract A computer-based expert system for the evaluation of research and development projects is presented. The system was developed for The Ministry of Science and Technology of the Republic of Slovenia and in the process of evaluation and selection of projects submitted to the annual competition for funds. The system is based on an adapted portfolio matrix that determines the position of each project with respect to its contents and feasibility. The aggregation of these criteria is carried out by a qualitative multi-attribute decision model that was developed using an expert system shell: DEX. The model consists of a tree of criteria, supplemented by if-then rules. In addition to describing these components, the paper presents and discusses a practical application of the system.

Evaluating options by combined qualitative and quantitative methods

Abstract This paper addresses two problems related to qualitative decision making: option ranking and non-sensitivity to small differences between options. In general, only a partial order of options can be established by a qualitative model, which might be insufficient particularly when the number of options is large. A qualitative model is also incapable of discriminating between slightly different options. In this paper, a solution is proposed that is based on an automatic construction of a quantitative evaluation model from the qualitative one. In addition to a qualitative class, a quantitative utility is obtained for each option, which is used to rank options within classes and to reflect the sensitivity to small differences between options.

Modelling impacts of cropping systems: Demands and solutions for DEX methodology

Decision modelling of diverse groups of problems makes different requirements to the modelling methodologies and software. We present an actual decision problem and the required characteristics of corresponding decision models. The problem is from agronomy and addresses the ecological and economic impacts of cropping systems, with the focus on the differences between cropping systems with conventional crops and the ones with genetically modified crops. We describe the extensions of an existing DEX qualitative multi-attribute modelling methodology, which were made to cope with the challenges of the problem. The extensions address general hierarchical structures, probabilistic utility functions and numerical values of basic attributes. A new, freely available software tool called proDEX was implemented to support the extended methodology. In this paper we describe the problem of cropping system assessment, propose methodological extensions to DEX, and present the implementation of proDEX.

Knowledge engineering techniques for utility identification

Abstract In this paper multiattribute decision making is discussed in terms of decision-making knowledge. Special emphasis is on identification (measurement) and verification of utility functions, and their use for evaluation of alternatives and explanation of evaluation results. Axiomatic and direct approach in utility theory are compared to the approach based on inductive learning techniques which are known from the field of artificial intelligence. Alternatives or their parts with the known utility are taken as learning examples in order to construct utility (function) knowledge. This approach is supported by a special expert system shell for utility knowledge modelling. It is implemented on a personal computer as a part of DECMAK system.

proDEX – A DSS tool for environmental decision-making

Abstract Environmental concepts are becoming common and ever more important parts of decision support models, which are a vital part of decision support systems. proDEX is a software tool for use of decision support models that are based on extended DEX methodology for qualitative multi-criteria decision modelling and support. The supported modelling methodology and the software features are adjusted to environmental modelling needs.

Constructing Intermediate Concepts by Decomposition of Real Functions

In learning from examples it is often useful to expand an attribute-vector representation by intermediate concepts. The usual advantage of such structuring of the learning problem is that it makes the learning easier and improves the comprehensibility of induced descriptions. In this paper, we develop a technique for discovering useful intermediate concepts when both the class and the attributes are real-valued. The technique is based on a decomposition method originally developed for the design of switching circuits and recently extended to handle incompletely specified multi-valued functions. It was also applied to machine learning tasks. In this paper, we introduce modifications, needed to decompose real functions and to present them in symbolic form. The method is evaluated on a number of test functions. The results show that the method correctly decomposes fairly complex functions. The decomposition hierarchy does not depend on a given repertoir of basic functions (background knowledge).