Search

The Third International Conference on Data Engineering was held in Los Angeles, California on February 2 to 6, 1987. The conference was designed as an international forum for bringing together researchers, developers, managers, strategic planners and other users with an interest in the research, design and development of data engineering methodologies, strategies and systems. Its scope includes computer science, artificial intelligence, electrical engineering, and computer engineering. It featured papers from all major areas of data engineering, including database design and modeling, performance evaluation, algorithms, integrity, security, fault tolerance, query language, artificial intelligence approaches, knowledge bases, database machines, distributed databases and data engineering applications.

The generalization error bounds found by current error models using the number of effective parameters of a classifier and the number of training samples are usually very loose. These bounds are intended for the entire input space. However, support vector machine (SVM), radial basis function neural network (RBFNN), and multilayer perceptron neural network (MLPNN) are local learning machines for solving problems and treat unseen samples near the training samples to be more important. In this paper, we propose a localized generalization error model which bounds from above the generalization error within a neighborhood of the training samples using stochastic sensitivity measure. It is then used to develop an architecture selection technique for a classifier with maximal coverage of unseen samples by specifying a generalization error threshold. Experiments using 17 University of California at Irvine (UCI) data sets show that, in comparison with cross validation (CV), sequential learning, and two other ad hoc methods, our technique consistently yields the best testing classification accuracy with fewer hidden neurons and less training time. [ABSTRACT FROM AUTHOR]

The Irvine School refers to an intellectual perspective on information technology in complex organizational settings that emerged at the University of California in Irvine, California (UCI), over the last three decades of the 20th century. In many ways, the rise of the Irvine School was synonymous with the rise of researcher Rob Kling's influence on the international community of scholars who would eventually form what Kling called social informatics. This article reflects on Kling's role in the creation of the Irvine School. Kenneth Kraemer, an architect and city planner by training, had joined the UCI Graduate School of Administration in 1967. He was one of the first scholars to begin careful empirical study of the effect of computerization in government agencies, and soon started a research program on information technology and local government at UCI's Public Policy Research Organization. In the early 1970s, researchers Jim Danziger and Rob Kling joined him in this endeavor. Danziger had been trained in political science and public administration. Kling had been trained in electrical engineering and computer science, and had done important work at the intersection of fuzzy logic and planning systems in the area of symbolic artificial intelligence.

Objective: To predict the true cost of developing and maintaining an electronic immunization registry, and to set the framework for developing future cost-effective and cost-benefit analysis.Data Sources/study Setting: Primary data collected at three immunization registries located in California, accounting for 90 percent of all immunization records in registries in the state during the study period.Study Design: A parametric cost analysis compared registry development and maintenance expenditures to registry performance requirements.Data Collection/extraction Methods: Data were collected at each registry through interviews, reviews of expenditure records, technical accomplishments development schedules, and immunization coverage rates.Principal Findings: The cost of building immunization registries is predictable and independent of the hardware/software combination employed. The effort requires four man-years of technical effort or approximately $250,000 in 1998 dollars. Costs for maintaining a registry were approximately $5,100 per end user per three-year period.Conclusions: There is a predictable cost structure for both developing and maintaining immunization registries. The cost structure can be used as a framework for examining the cost-effectiveness and cost-benefits of registries. The greatest factor effecting improvement in coverage rates was ongoing, user-based administrative investment. [ABSTRACT FROM AUTHOR]