David West

A comparison of SOM neural network and hierarchical clustering methods

Abstract Cluster analysis, the determination of natural subgroups in a data set, is an important statistical methodology that is used in many contexts. A major problem with hierarchical clustering methods used today is the tendency for classification errors to occur when the empirical data departs from the ideal conditions of compact isolated clusters. Many empirical data sets have structural imperfections that confound the identification of clusters. We use a Self Organizing Map (SOM) neural network clustering methodology and demonstrate that it is superior to the hierarchical clustering methods. The performance of the neural network and seven hierarchical clustering methods is tested on 252 data sets with various levels of imperfections that include data dispersion, outliers, irrelevant variables, and nonuniform cluster densities. The superior accuracy and robustness of the neural network can improve the effectiveness of decisions and research based on clustering messy empirical data.

Model selection for a medical diagnostic decision support system: a breast cancer detection case

There are a number of different quantitative models that can be used in a medical diagnostic decision support system (MDSS) including parametric methods (linear discriminant analysis or logistic regression), non-parametric models (K nearest neighbor, or kernel density) and several neural network models. The complexity of the diagnostic task is thought to be one of the prime determinants of model selection. Unfortunately, there is no theory available to guide model selection. Practitioners are left to either choose a favorite model or to test a small subset using cross validation methods. This paper illustrates the use of a self-organizing map (SOM) to guide model selection for a breast cancer MDSS. The topological ordering properties of the SOM are used to define targets for an ideal accuracy level similar to a Bayes optimal level. These targets can then be used in model selection, variable reduction, parameter determination, and to assess the adequacy of the clinical measurement system. These ideas are applied to a successful model selection for a real-world breast cancer database. Diagnostic accuracy results are reported for individual models, for ensembles of neural networks, and for stacked predictors.

Ensemble strategies for a medical diagnostic decision support system: A breast cancer diagnosis application

The model selection strategy is an important determinant of the performance and acceptance of a medical diagnostic decision support system based on supervised learning algorithms. This research investigates the potential of various selection strategies from a population of 24 classification models to form ensembles in order to increase the accuracy of decision support systems for the early detection and diagnosis of breast cancer. Our results suggest that ensembles formed from a diverse collection of models are generally more accurate than either pure-bagging ensembles (formed from a single model) or the selection of a ‘‘single best model.’’ We find that effective ensembles are formed from a small and selective subset of the population of available models with potential candidates identified by a multicriteria process that considers the properties of model generalization error, model instability, and the independence of model decisions relative to other ensemble members. � 2003 Elsevier B.V. All rights reserved.

Model selection for medical diagnosis decision support systems

In this paper, we examine the model section decision for a medical diagnostic decision support system (MDSS). Our purpose in doing this is to understand how model selection affects the accuracy of the decision support system. We explore two related research questions: (1) Do ensembles of models, acting as a single decision maker, perform more accurately than single models; and (2) How does model diversity affect the accuracy of the ensembles? Specifically, we compare 23 single models and bootstrap aggregating (i.e., bagging) models for their predictive abilities across five diverse medical data sets. We are able to reach important conclusions about our research objectives. Ensembles are more accurate than single models in their predictive ability. The best ensemble model achieves an error level significantly lower than the error of the best single model for four of the five medical applications analyzed. The magnitude of the error reduction ranges from 6.4% to 17.5%. Also, when designing an ensemble for an MDSS, the decision to diversify the model selection should be guided by the relationship between model instability and generalization error for the population of models under consideration.

Predictive modeling for wastewater applications: Linear and nonlinear approaches

This study compares the multi-period predictive ability of linear ARIMA models to nonlinear time delay neural network models in water quality applications. Comparisons are made for a variety of artificially generated nonlinear ARIMA data sets that simulate the characteristics of wastewater process variables and watershed variables, as well as two real-world wastewater data sets. While the time delay neural network model was more accurate for the two real-world wastewater data sets, the neural networks were not always more accurate than linear ARIMA for the artificial nonlinear data sets. In some cases of the artificial nonlinear data, where multi-period predictions are made, the linear ARIMA model provides a more accurate result than the time delay neural network. This study suggests that researchers and practitioners should carefully consider the nature and intended use of water quality data if choosing between neural networks and other statistical methods for wastewater process control or watershed environmental quality management.

IMPROVING DIAGNOSTIC ACCURACY USING A HIERARCHICAL NEURAL NETWORK TO MODEL DECISION SUBTASKS

A number of quantitative models including linear discriminant analysis, logistic regression, k nearest neighbor, kernel density, recursive partitioning, and neural networks are being used in medical diagnostic support systems to assist human decision-makers in disease diagnosis. This research investigates the decision accuracy of neural network models for the differential diagnosis of six erythematous-squamous diseases. Conditions where a hierarchical neural network model can increase diagnostic accuracy by partitioning the decision domain into subtasks that are easier to learn are specifically addressed. Self-organizing maps (SOM) are used to portray the 34 feature variables in a two dimensional plot that maintains topological ordering. The SOM identifies five inconsistent cases that are likely sources of error for the quantitative decision models; the lower bound for the diagnostic decision error based on five errors is 0.0140. The traditional application of the quantitative models cited above results in diagnostic error levels substantially greater than this target level. A two-stage hierarchical neural network is designed by combining a multilayer perceptron first stage and a mixture-of-experts second stage. The second stage mixture-of-experts neural network learns a subtask of the diagnostic decision, the discrimination between seborrheic dermatitis and pityriasis rosea. The diagnostic accuracy of the two stage neural network approaches the target performance established from the SOM with an error rate of 0.0159.

An improved neural classification network for the two-group problem

Abstract In this paper we present the neural network model known as the mixture-of-experts (MOE) and determine its accuracy and its robustness. We do this by comparing the classification accuracy of MOE, backpropagation neural network (BPN), Fisher’s discriminant analysis, logistics regression, k nearest neighbor, and the kernel density on five real-world two-group data sets. Our results lead to three major conclusions: (1) the MOE network architecture is more accurate than BPN; (2) MOE tends to be more accurate than the parametric and non-parametric methods investigated; (3) MOE is a far more robust classifier than the other methods for the two-group problem. Scope and purpose The two-group classification problem is the assignment of objects to one of two predetermined groups. These classification decisions are routinely made in business, health care, and government. Classification errors result when an object is assigned to the wrong group. A healthy patient might be diagnosed with cancer, or a patient suffering from cancer might be diagnosed as healthy. A firm assigned the wrong bond rating will incur additional cost of capital of millions of dollars. Unfortunately, there is no classification method that is the “best” for all data sets. Some classification methods will do well for a given data set only to perform poorly on another. What is needed, therefore, is a robust methodology that classifies accurately across a wide range of two-group data sets. Researchers and practitioners have turned to neural classification models in a quest for a robust method. The purpose of this research is to investigate the accuracy and robust nature of the mixture-of-experts neural network (MOE). Our research demonstrates that MOE performs very well against back-propagation neural networks and four traditional statistical classification procedures based upon results from five diverse real world data sets. We conclude that the mixture-of-experts neural network may be the robust classification scheme being sought for use with the two-group problem.

Overbooking Increases Patient Access at East Carolina University's Student Health Services Clinic

The health-care clinic presented in this study experienced significant numbers of patients who failed to arrive for their scheduled appointments (no-shows). The cost of reducing patient access at this clinic because of no-shows is estimated to exceed

Transfer function modeling of processes with dynamic inputs

400,000 annually. An interdisciplinary quality-improvement team developed a novel health-care overbooking model that includes the effects of employee burnout. This model estimates the nonlinear nature of the costs associated with medical-provider burnout caused by overbooked appointments that exceed clinic capacity. Several key East Carolina University clinical staff members had been skeptical about the value of overbooking. The model was instrumental in convincing them that implementing an overbooking process would benefit patients and the organization. The clinic, which subsequently implemented such a process, attributes a savings of

Control of complex manufacturing processes: a comparison of SPC methods with a radial basis function neural network

95,000 per semester to the initiative.