Decision Tree Methodology Research Articles

Applying decision tree methodology for rules extraction under cognitive constraints

Within this paper the initialization phase of the cognitive management for anthropocentric production systems (COMAPS) project (BRITE-EURAM BE 96-3941) is presented. The entire COMAPS-system is an online decision support system to control dynamically changing industrial production processes. This paper focuses on the adoption of production rules produced by decision tree techniques to the cognitive model of expert's control practice. The applied algorithms to obtain a quasi-optimal rule set as a model of the intentional representation of the expert knowledge is presented and tested on real world applications (printed circuit board, foil and brake pad production). The results of one application are shown.

An approach to demonstrating cost-effectiveness of diagnostic imaging modalities in Australia illustrated by positron emission tomography.

The aim of this study was to develop a framework in which the cost-effectiveness of new imaging technologies could be evaluated using data from other countries, while assessing the impact that any differences between the study populations and Australia may have upon the results. Publications reporting the cost-effectiveness or therapeutic impact of positron emission tomography (PET) were re-worked using Australian cost structures. PET was assigned a cost of $950. The effects of potential differences between the populations studied and the Australian population were evaluated by applying sensitivity analysis to those publications that describe decision tree methodology. The parameters included in the sensitivity analysis were disease prevalence and specificity of PET. The Australian cost savings per patient examined by PET were $505.50-$912.41 for investigation of solitary pulmonary nodules, $34.65-$360.03 for lung cancer staging, $550.08 for axillary staging of breast cancer, $230.75-$2301.27 for assessment of recurrent colorectal cancer and $300.24-$2069.65 for assessment of myocardial viability. Significant differences in disease prevalence and PET specificity could occur while the cost-effectiveness of PET was preserved. Decision tree sensitivity analysis can demonstrate the cost-effectiveness of diagnostic imaging modalities in Australia and provides indications that PET is cost-effective for a range of clinical indications.

Discovering Decision Knowledge from Web Log Portfolio for Managing Classroom Processes by Applying Decision Tree and Data Cube Technology

In conventional classrooms, teachers attempt to enhance instruction by monitoring students' learning processes and analyzing their performances by paper records and observation. Similarly, distance learning systems on the Web should be designed to record students' behaviors to assist teachers in assessing performance and making decisions related to curriculum. Recent developments in Web server systems can record the students' access to the learning systems in Web logs. Information processing analysis on the historical classroom processes can help teachers to develop knowledge for applying proper teaching strategies according to available information in Web logs. However, teachers cannot easily infer the pedagogical meaning of Web logs and discover the pedagogical rules of students' behavior patterns in the Web logs to refine teaching strategies. Therefore, to use Web logs for pedagogical purposes, this article adopts decision tree and data cube information processing methodologies to observe students' behaviors and discover the pedagogical rules on students' learning performance from Web logs. The architecture and guidelines of utilizing the data cube and decision tree methodologies for pedagogical purposes are also presented. Consequently, teachers can efficiently estimate and explain the effectiveness of pedagogical strategies, ultimately improving instruction with decision tree and data cube software.

Quantitative decision making for cued land reconnaissance: a defence case study

AbstractHelping commanders make more informed decisions is one of the most important roles played by defence operations analysts. This paper presents a defence case study of quantitative decision making. It develops a decision model for cued land reconnaissance missions that treats detecting, identifying, and destroying a hostile target to protect an asset as an integral process. The paper investigates issues such as whether an information assessment system, which provides improved knowledge about the nature of the cue, is necessary and what kind of reconnaissance platforms should be used under various conditions. Using a decision tree and evaluating all possible outcomes, the model derives the best course of action based on expected values. Then serials of sensitivity analysis are carried out to account for uncertainties of the scenarios, and their implications are explored. This illustrates that decision tree methodology can be used to analyse complex sequential decision problems in real military situations, and uncertainties can be taken into account in the decision making process.

Quantitative decision making for cued land reconnaissance: a defence case study

Helping commanders make more informed decisions is one of the most important roles played by defence operations analysts. This paper presents a defence case study of quantitative decision making. It develops a decision model for cued land reconnaissance missions that treats detecting, identifying, and destroying a hostile target to protect an asset as an integral process. The paper investigates issues such as whether an information assessment system, which provides improved knowledge about the nature of the cue, is necessary and what kind of reconnaissance platforms should be used under various conditions. Using a decision tree and evaluating all possible outcomes, the model derives the best course of action based on expected values. Then serials of sensitivity analysis are carried out to account for uncertainties of the scenarios, and their implications are explored. This illustrates that decision tree methodology can be used to analyse complex sequential decision problems in real military situations, and uncertainties can be taken into account in the decision making process.

CLOSURE PLAN EVALUATION FOR RISK OF ACID ROCK DRAINAGE

Control of acid rock drainage (ARD) is a long-term issue for many mine sites and is often a primary objective of remediation efforts. Some sites continue to require monitoring and management of ARD long after mine operation has ceased and closure is complete. In New Zealand, an innovative and quantitative approach was applied to evaluate the expected risk of ARD after implementation of the closure plan for the Golden Cross Mine. In addition, this future risk was compared to current operating conditions to provide an estimate of the reduction in risk provided by the remediation activities. This approach was useful to both the mine proponent and the regulatory agencies in assessing the effectiveness of the existing closure plan and providing focus on the components of greatest risk. Mine components remaining on site after closure that could potentially generate ARD under various failure scenarios were identified and evaluated. These components included the tailings decant pond, waste rock, stockpiles, open pit mine and water treatment systems. For each component, a series of initiating events and failure scenarios were identified, and a decision tree methodology was utilized to estimate the probability of ARD generation for both current and closure conditions. Due to themore » implementation of closure plans designed to minimize or eliminate ARD through regarding, construction of engineered covers and water management designs, the risk of ARD generation will be significantly reduced over time.« less

Reengineering claims processing using probabilistic inductive learning

With health care costs in the United States skyrocketing, and $.25 of every health care dollar being spent on systems and claims administration, technological advances such as electronic claims filing are being advocated as cost-reducing measures. These improvements alone, however, will not significantly reduce costs unless they are accompanied by revisions in the entire claims processing system. This study explores the reliability and utility of probabilistic inductive learning (PrIL), a statistically enhanced decision tree algorithm, for improving the decision-making process at the New York State Workers' Compensation Board (WCB). Results indicate that the PrIL algorithm is favorably comparable to both the purely statistical and the classical decision tree methodologies, with the added advantages of easy to understand rules and user-defined reliability measures for each of those rules. Given the appropriate information regarding the relative value of correct and incorrect classification of cases in the WCB system, PrIL can be used to accurately assist in the decision making process in terms of reducing cost, predicting and enhancing quality and case outcomes in managed care practices.

Assessing school readiness for a practice arrangement using decision tree methodology.

Decision-tree methodology provides a systematic approach for schools to use in examining their own situation to determine if a faculty practice arrangement is appropriate for their particular school at this particular time. Questions in the decision tree address mission, faculty interest, administrative support, a practice plan, and practice models. Specific questions in the decision tree include the following: Does the institutional mission include service/practice? Does the college/school mission include practice? Does the college/school have a critical mass of faculty interested in practice? Does college/school administration support practice? Does a critical mass of faculty desire a faculty practice plan? Can faculty and administration work together to produce a mutually beneficial practice arrangement? Do faculty and administration desire to operate their own practice arrangement? Is there an identified need for nursing services that can be met by the college/school? Does the internal and external environment support the college/school operating a practice? Do internal or external organizations desire the talents/expertise of faculty members? Schools frequently evaluate two models, the Integration/Nursing Center Model and the Organizational Contracting/Entrepreneurship Model, for their specific risks and benefits. Critical to the success of any practice arrangement is that there be congruency between the school's allocation of human resources and its reward system.

Decision Tree Method for the Classification of Chemical Pollutants: Incorporation of Across-Chemical Variability and Within-Chemical Uncertainty

We have developed a decision tree methodology for the classification of chemicals by estimates of potential human exposure. The steps involved in the construction of a decision tree are as follows. Monte Carlo simulations are conducted by randomly sampling chemical and environmental properties, whose range of values represents the variability of parameters across a defined set of chemicals and environmental conditions. The tree structure is then defined by a series of constraints placed on the various chemical and environmental properties using the Clas sification and Regression Tree Algorithm (CART). Each node of the tree is associated with a human exposure value and is considered a bin, which classifies chemicals whose properties are consistent with those parametric constraints associated with the particular node. In addition to being associated with parametric constraints, each bin or tree node is associated with a human exposure level. In this manner, the tree structure functions as a template from wh...

Decision trees for fast security assessment of autonomous power systems with a large penetration from renewables

This paper presents the application of the decision tree methodology to the dynamic security assessment of a medium size autonomous power system with a large penetration from renewable energy sources. The effect of various parameters is investigated and the perspectives of the integration of this scheme in the advanced control system of the Greek island of Lemnos are discussed.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Design of multicategory multifeature split decision trees using perceptron learning

The decision tree methodology is an important nonparametric technique for building classifiers from a set of training examples. Most of the existing top-down decision tree design methods make use of single feature splits at successive stages of the tree design. While computationally attractive, single feature splits generally lead to large trees and inferior performance. This paper presents a new top-down decision tree design method that generates compact trees of superior performance by using multifeature splits in place of single feature splits at successive stages of the tree development. The multifeature splits in the proposed method are obtained by combining the concept of information measure of a partition with perceptron learning. Several decision tree induction results for a broad range of classification problems are presented to demonstrate the strengths of the proposed decision tree design methods.

Decision tree approach to power systems security assessment

An overview of the general decision tree approach to power system security assessment is presented. The general decision tree methodology is outlined, modifications proposed in the context of transient stability assessment are embedded, and further refinements are considered. The approach is then suitably tailored to handle other specifics of power systems security, relating to both preventive and emergency voltage control, in addition to transient stability. Trees are accordingly built in these various application domains, and their salient features are explored, assessed, compared. Among attractive aspects of the trees, we mention their ability to uncover the intrinsic mechanism governing physical processes, and to provide a clear description in terms of tractable system parameters. Further, their dual, ‘attribute space’ representation is shown to have complementary possibilities for more refined analysis and, in addition, for sensitivity assessment and control. Overall, the approach is characterized by its great flexibility with respect to tree structure, types of physical parameters driving the phenomena, and classes of potential applications. To illustrate and support the developments, real-world examples are reported, relating to transient and voltage stability issues simulated on several power systems.

Decision tree approaches to voltage security assessment

A decision tree methodology is proposed to tackle voltage security concerns under two distinct facets, namely preventive-wise and emergency-wise. Decision tree methodology in general consists of inferring information from a ‘learning set of states’, and of compressing while hierarchically organising this information in a tree fashion, in terms of salient parameters automatically selected by the method at its nodes. Subsequently, the tree may readily assess any already known or new state. For preventive voltage stability purposes, trees are built for a given contingency and the system robustness is assessed in terms of precontingency parameters. For emergency control purposes, on the other hand, each tree is built for a set of contingencies and the system robustness is assessed in terms of parameters observed just after a contingency occurrence. The proposed approaches are illustrated by means of two simulation studies: one carried out on the Brittany region of the EHV French system and the other on an academic type example. These studies explore feasibility aspects, highlight interesting features and suggest that the proposed approaches have indeed potential for real world applications.

Generalized decision trees: methodology and applications

The method of decision trees is a widely accepted decision tool for multistage decision problems. The classical decision tree methodology assumes that the future outcome, following a decision at a given stage, is a discrete random variable with known probability function. Moreover, with each decision, the resultant cash flows and their timings are assumed to be deterministic and constants. However, in many real life problems in a dynamic environment, in addition to the probabilistic nature of outcomes, there is also randomness in the size of cash flows as well as their timings. By generalized decision trees (GDTs) we mean decision trees with cash flows as well as their timings being random variables. Furthermore, the time between two consecutive decision points is also allowed to be a random variable in the dynamic decision process. The method which is widely used for these types of decision problems is the Monte Carlo simulation. In this paper we present an analytical method for obtaining optimal decisions in generalized decision trees. Applications of generalized decision trees to optimal recall time of defective products (e.g. cars), optimal software release time and decision of automation in a manufacturing environment are also presented.

Inductive inference applied to on-line transient stability assessment of electric power systems

A decision tree methodology is proposed for the purpose of predicting the robustness of a power system in the occurrence of severe disturbances, and of discovering appropriate control actions, whenever needed. The decision trees are built off-line and used on-line. Their building calls upon an inductive inference method, which automatically extracts the relevant information from large bodies of simulation data and expresses it in terms of the system controllable variables. The trees provide a clear understanding of the intricate mechanisms of transient stability, and means to control. This paper focuses on fundamental issues relative to the inductive inference method. A real world power system comprising 14 generators, 112 branches and 92 nodes is used to illustrate it.

Selection of a manufacturing process for robots

A decision tree methodology for selecting the use of robots in a manufacturing system has been developed for sheet metal work.