Jean E. Weber

An Interactive Multiple Objective Decision-Making Aid Using Nonlinear Goal Programming

A nonlinear goal programming approach is embedded within an interactive framework allowing the decision maker to direct the algorithm’s search for a satisfactory alternative. Particular consideration has been given to the psychological assumptions required in this approach, especially the Gestalt nature of perception, the information-dependent nature of acceptability, and the serial aspects of the selection process.

Quantitative study of ductal breast cancer--patient targeted prognosis: an exploration of case base reasoning.

Current analytic methodologies allow the extraction, even from small tumor masses, of extensive information on the biologic characteristics of malignant lesions, such as tumor aggressivity, metastatic potential, drug resistance, and host interactions. Clinical practice now offers a wide range of therapeutic strategies. Information technological advances offer the opportunity to refer to very large data bases of patient anamnestic data, response to treatment and clinical outcome. There is a need to formulate therapy and prognosis for each individual case. Case based reasoning is a knowledge based methodology where the outcome for complex situations can be predicted by referring to a large data base of cases of known outcomes. The preliminary data obtained from this study suggest that case based reasoning may offer a promising approach to individual targeted prognosis.

Organization of a knowledge base by Q-analysis

Q-analysis is used in organizing a histopathological knowledge base which is a component of the diagnostic expert system at the University of Arizona. This expert system has three subsystems or modules: the first module guides the dynamic reconfiguration of the computer system; the second module guides scene decomposition and the extraction of diagnostic information; the third module uses a rule-based system to obtain a diagnostic assessment. A specific example of the usefulness of Q-analysis is given in the context of the third module; the data represent expert opinion concerning diagnosis of colonic cancer and are summarized in a matrix representing four diagnostic categories and nineteen diagnostic clues. The example shows that Q-analysis may be helpful to diagnosticians in defining and explaining the process they use in arriving at a diagnosis and in using this information as a basis for structuring the knowledge base.

Construction of the knowledge file for an image understanding system

To enable an image understanding system to provide an automated interpretation of diagnostic imagery it must have access to all of the concepts, procedures and methods used by human experts. The paper describes information elicitation from experts of different domains and the construction of a knowledge file. Uncertainty management is based on Bayesian belief network methods.

Performance of the Durbin-Watson Test and WLS Estimation When the Disturbance Term Includes Serial Dependence in Addition to First-Order Autocorrelation

Abstract Monte Carlo simulation is used to study the power of the Durbin-Watson test and the properties of the corresponding weighted least squares (WLS) estimates when there is serial correlation in the disturbance term, in addition to first-order autocorrelation. The results indicate that the Durbin-Watson test detects first-order autocorrelation, even when other forms of serial dependence are also present. However, routine use of WLS estimation when the Durbin-Watson test is significant may result in inaccurate and inefficient parameter estimates. Therefore, this procedure should be used with caution unless there is a priori knowledge concerning the nature of any serial dependence in the disturbance terms.

Ultrastructural morphometric analysis of papillary neoplasms : biological and diagnostic relevance

Ten papillary adenocarcinomas of thyroid origin (P-Thy), ten papillary adenocarcinomas of ovarian origin (P-Ov), and eight papillary neoplasms of non-thyroid/non-ovarian origin (P-Other) were morphometrically compared using 19 distinct quantitative nuclear and nucleolar parameters as a database for diagnosis. The selected cases consisted of 16 primary and 12 metastatic neoplasms. It was determined that the P-Thy group had a significantly smaller nucleolar area (NuA) and nucleolar perimeter (NuP), and smaller SDs of nuclear area (NA), NuA, and NuP compared with the P-Ov and P-Other groups (P less than .05). The P-Ov group had a significantly smaller SD of NA compared with the P-Other group (P less than .05). The P-Ov group exhibited the greatest variability among the papillary neoplasms. Linear regression analysis indicated that in the P-Thy group alone there was a significant correlation between mean nuclear form factor (4 pi A/P2) and mean NuA (r = -.82; P less than .01), and mean NP and mean NuA (r = +.77; P less than .01). Linear regression analysis also indicated that in the P-Ov group alone, there was a significant correlation between mean NA and mean NuA (r = +.75; P less than .02). Morphometric domains were established using statistically significant sets of variables that distinguished between the groups. The application of three-dimensional computerized cluster analysis techniques indicated that the P-Thy group consistently had the smallest morphometric domains. It was concluded that ultrastructural morphometric analysis of papillary neoplasms has diagnostic potential and reveals interesting biological relationships among distinct nuclear features in the different groups of neoplasms.

Statistical identification of subpopulations for flow cytometric data

Identification of cell subpopulations is of interest in the context of both flow cytometry and image analysis. Flow cytometry makes it possible to examine very large cell populations rapidly and to record a measurement vector for each cell. Thus flow cytometry, as a methodology, is ideally suited to the identification of small subpopulations of cells, which is particularly of interest in analyzing lymphoid cell populations.

Guided data reduction for flow cytometry

A statistical procedure for identifying subpopulations in large samples is discussed in the context of the analysis of flow cytometric data. The PINDEX algorithm is proposed for analyzing such multimodal data. PINDEX is a clustering algorithm which considers the covariance structure of clusters. When such an algorithm is run in a brute force manner, the number of runs required for convergence may involve computation which, for large samples, is prohibitive. A solution to this problem using starting centroids chosen by an expert system is proposed. Sample runs were made using data sets for lymphoid cells; these data sets are of modest dimensionality and contain four subpopulations. The results demonstrate that substantive economy of processing can be achieved by having the expert system provide start-up cluster centroids close to the final coordinates.<<ETX>>

The Contribution of Long-Term Records of Hydrologic Extremes to Risk Analyses

We discuss the advantages of lengthening the records of extreme floods and droughts by geological and botanical data and propose a methodology to incorporate such data into risk analyses. By themselves, such data provide information on the most extreme events that actually occurred in a drainage basin or region during the last several centuries to millennia. Through the adaptation of existing methodologies, these records serve as prior information within the Bayesian framework and substantially improve the estimated parameters of the probability distribution functions of extreme phenomena.

Conceptual learning in an expert system for histopathologic diagnosis

Autonomous learning modules in a diagnostic expert system serve to reveal distributional properties of the diagnostic clue values. This leads to more exhaustive utilization of the collected information. It also results in matching the design granularity for the diagnostic discrimination to the true structure of the diagnostic data. Additional conceptual entities augment the knowledge base.