Scott P. Stevens

An Empirical Evaluation of Probability Estimation with Neural Networks

Recent interest in neural networks by researchers across a wide spectrum of disciplines has provided convincing evidence of their ability to address classification problems. In this article, we consider the issue of evaluating the predictive capability of neural networks when the output values are to be treated as probabilities. We propose the use of a variant of a chi-square statistic, based on the Hosmer–Lemeshow statistic from logistic regression, to measure the goodness-of-fit of neural network models for two-group membership problems. Through experimentation with a large real-world database, we demonstrate the application of this statistic, and examine the effects of varying the number of nodes in the hidden layer on its value. Our empirical results suggest that this statistic can be very useful in identifying significant differences in the probability estimation accuracy of neural network models.

Using neural networks for trauma outcome evaluation

Abstract The development of effective methods for predicting the survival or mortality of patients is a major focus of trauma research. We address this problem by applying neural network modelling, using as inputs established descriptors of physiologic and anatomic injury severity. We use two different “backpropagation” techniques with a variety of training strategies and examine the importance of training and test set composition, performance evaluation criteria, and interpretation of network outputs. The results of our study indicate that neural networks show significant promise for trauma patient outcome evaluation.

Automatic Data Collection Systems: Observed Benefits and Problems

The current goal of many US firms is to become the highest quality and lowest cost leaders in their markets. Achieving this goal requires major changes in how products and services are produced. Many systems have been developed to assist firms in accomplishing this goal; they include such diverse areas as Just‐in‐Time manufacturing (JIT), Material Resource Planning II (MRP II), Automatic Storage and Retrieval Systems (ASRS), Electronic Data Interchange (EDI), CAD/CAM, Computer Integrated Manufacturing (CIM), and Flexible Manufacturing Systems (FMS). One key element in the success of these systems is that accurate data concerning plant operations must be immediately available to those who need it. More and more frequently, this requirement is being addressed through the use of automatic data collection systems (ADC). Seeks to provide insights into the application of ADC systems in US industries. Considers the type of data collection system installed, the implementation and operational problems encountered,...

Teaching Use of Binary Variables in Integer Linear Programs: Formulating Logical Conditions

Binary variables are often needed in linear programming models to indicate whether particular alternatives should be implemented and to impose logical relations among decisions. However, it is usually not obvious to students how to use binary variables to transform conditional statements of logic into linear relations. We propose to address this difficulty with a simple two-step approach. It provides rules for decomposing a conditional requirement into a group of elementary implications and then translating each of these into linear constraints. Pre- and post-test results from a sample of undergraduate business students are presented to support the effectiveness of this pedagogical approach. Supplemental files available online at https://doi.org/10.1287/ited.2017.0177.

Lessons from the Virtual Campus: The Life and Death of a Second Life University

Second Life is a virtual world where many universities have established an educational presence. Hoping to serve as a useful example for others that are considering the creation of a virtual campus, this article follows how the Second Life campus of James Madison University was created in 2008. It details the nontraditional ways that it was used by faculty and students for museum studies, student projects, and in-world conferencing. It also reviews the major reasons why the project failed in April 2012 after four years of successful use, concluding with thoughts on best practices for Second Life university campuses.

Computer Ethics of University Students: An International Exploratory Study

Compares attitudes regarding the ethical use of computers in samples of business students from universities in the United States and Australia. Finds that the two groups are statistically significantly different on 17 of the 26 ethics questions. Results show that significant differences in attitudes towards property and privacy issues exist; attitudes toward ethics in employee‐employer relations were not significantly different. In general, students in Australia tend to take ethical positions supporting greater freedom of action for computer users, while students in the United States take more restrictive positions.

A simple stochastic model for computer terminal availability

Queueing models have been widely applied in the analysis of computer performance, as well as in the study of systems in which service is rendered to human customers by multiple servers. Rarely, however, have queueing theory techniques been applied to the problem of providing computing services to the human beings who require them.We develop a simple queueing model for the availability of computer terminals in an academic computing laboratory, then compare the statistics predicted by the model to those actually observed in several laboratories. Among the statistics considered is the likelihood that a new arrival to the laboratory finds no free terminals, a quantity which is comparable to standard performance measures for communications equipment. The model also generates measures of resource availability which may be analyzed dynamically. Examples of the use of this model are provided.