This paper proposes a model to analyze asymmetric matrix of proximity, of which form is expressed in terms of distance and additive parameters. Two versions of the model are treated, one for matrix with diagonal entries, the other for matrix with diagonal elements undefined. For infallible data, necessary and sufficient conditions are considered, and algebraic solutions are suggested. To deal with fallible data, we propose scaling procedures based on least squares criteria. Applications are provided for demonstration.

The Freiburg University Hospital (2 000 beds, 530 000 outpatient visits per year) has developed the second generation of its computerized information system. Mainframe computers with 560 terminals (display terminals, printers, PCs, card imprinters) and an ISDN based PBX System, especially for transmission of findings, are used centrally. Decentrally organized department systems were bought or developed for the Radiology, Hygiene, Central Laboratory, Pathology etc. and connected with the central systems. Most of the department systems are based on mini computers but there are also PC-LAN e.g. for the Tumor Center, Surgery and Neurophysiology. We designed a new relational patient data base with some deviations from the normal forms and a special error tolerant patient identification. The non patient oriented administrative applications are mainly based on commercially available SAP software. The most sophisticated subsystem is the PACS which stores and transmits radiological pictures.

Summary

Dynamic exponential family regression provides a framework for nonlinear regression analysis with time dependent parametersβ₀,β₁, …,β_t, …, dimβ_t=p. In addition to the familiar conditionally Gaussian model, it covers e.g. models for categorical or counted responses. Parameters can be estimated by extended Kalman filtering and smoothing. In this paper, further algorithms are presented. They are derived from posterior mode estimation of the whole parameter vector (β′₀, …,β′_t) by Gauss-Newton resp. Fisher scoring iterations. Factorizing the information matrix into block-bidiagonal matrices, algorithms can be given in a forward-backward recursive form where only inverses of “small”p×p-matrices occur. Approximate error covariance matrices are obtained by an inversion formula for the information matrix, which is explicit up top×p-matrices.

In the process of exploring a method of presenting medical knowledge in a mainly graphical way for an “intelligent” textbook, a system has been implemented on a graphics workstation. The intention was not to build a consultation system, but rather to improve access to knowledge by visualizing complex relationships: the main feature is graphical browsing through a semantic net displayed on the screen. Nodes represent diseases, pathophysiological states, and findings; detailed information about the nodes may be displayed via pull-down menues. The main links represent “is-a”, “causes” (with variants like “causes with delay” or “predisposes-to”), “is-associated-with”, and “excludes” relationships. The first application domain of the system is gastroenterology. To date, 100 diseases, 100 pathophysiological states, and 400 findings have been entered.

A Bayesian solution is given to the problem of making inferences about an unknown number of structural changes in a sequence of observations. Inferences are based on the posterior distribution of the number of change points and on the posterior probabilities of possible change points. Detailed analyses are given for binomial data and some regression problems, and numerical illustrations are provided. In addition, an approximation procedure to compute the posterior probabilities is presented.

This paper describes the new concept of an EMG-System. A transputer-based hardware allows the analysis of EMG-signals with different methods at the same time. Additionally an expert system can run on the same device, so that a fast diagnosis support is possible.