Search

CONTEXT: Insulin resistance (IR) is a risk factor for type 2 diabetes, diabetic kidney disease, cardiovascular disease and nonalcoholic steatohepatitis. Biliopancreatic diversion (BPD) is the most effective form of bariatric surgery for improving insulin sensitivity. OBJECTIVE: To identify plasma proteins correlating with the early restoration of insulin sensitivity after BPD. DESIGN: Prospective single-center study including 20 insulin-resistant men with morbid obesity scheduled for BPD. Patient characteristics and blood samples were repeatedly collected from baseline up to 4 weeks postsurgery. IR was assessed by homeostatic model assessment for insulin resistance (HOMA-IR), Matsuda Index, and by studying metabolic profiles during meal tolerance tests. Unbiased proteomic analysis was performed to identify plasma proteins altered by BPD. Detailed plasma profiles were made on a selected set of proteins by targeted multiple reaction monitoring mass spectrometry (MRM/MS). Changes in plasma proteome were evaluated in relation to metabolic and inflammatory changes. RESULTS: BPD resulted in improved insulin sensitivity and reduced body weight. Proteomic analysis identified 29 proteins that changed following BPD. Changes in plasma levels of afamin, apolipoprotein A-IV (ApoA4), and apolipoprotein A-II (ApoA2) correlated significantly with changes in IR. CONCLUSION: Circulating levels of afamin, ApoA4, and ApoA2 were associated with and may contribute to the rapid improvement in insulin sensitivity after BPD.

An appraisal of the quality of methodology and software for performing cost uncertainty analysis will be given. It will be argued that the standard practice of developing a 'most likely cost' and then generating an uncertainty distribution associated with that figure is a consequence of historical precedent and places the cart before the horse. It will also be argued that methodology for encoding subjective probability distributions are out of date, being, for the most part, modifications of the PERT techniques of the sixties. Additionally, methodology for processing elemental uncertainty assessments into uncertainty distributions over higher order structures (e.g. subsystem CER's or total system cost) often ignores important sources of variability, and extent computer software designed to implement cost uncertainty analysis methodology is often poorly written and, in some cases, may lead to conclusions which are inconsistent with the inputs. This is a paper presented at a cost analysis symposium., Presented at the Annual Department of Defense Cost Analysis Symposium (19th) held at Leesburg, Virginia, on 17-20 September 1985.

Given the current emphasis on improving all aspect of cost analysis and the primacy of cost uncertainty in the cost estimation process, it is time to assess the quality of the state-of-the-art and look toward improving techniques where needed. An argument supporting the primacy of cost uncertainty analysis and a critical assessment of existing techniques was given. In this paper, we examine some efforts to apply new results and insights, both mathematical and psychological, to the problems of cost uncertainty analysis. The paper will be treat some new ideas pertaining to probability elicitation/ encoding using interactive software and indirect assessment of the uncertainty associated with a CER., Presented at the Annual Department of Defense Cost Analysis Symposium (19th), Leesburg, VA, 17-20 Sep 1985.

This paper discusses the basket method of sampling and estimation in finite populations in the framework of the traditional design based approach (randomization) and the model oriented approach (superpopulation). It is shown that the basket method incorporates the desirable features of both approaches. Its properties are studied. (Author), Supersedes rept. no. N129.

This Final Report lists project personnel, catalogues 158 Technical Reports by Number, title, and author and cites bibliographic reference for the 71 reports which have been published in professional literature during the course of the contract. Keywords: Mathematical sources; Naval research. (Author)

This dissertation explores the distributional properties of commonly used statistics developed in the course of empirical model building. A review of some of the more noteworthy efforts to investigate the distribution of the coefficient of determination of determination, R-squared, in the best subset regression is given. To overcome the shortcomings of these results, a permutation test based on Fisher's randomization test is developed to provide a practical basis for assessing the statistical significance of a regression in such situations. An investigation is made into the distributional properties of the multiple correlation coefficient in the choice of a transformation of the dependent variable. The study investigates the possibility that pedestrian use of transformations, such as y* = (y+c) to the p, may lead to an inflationary effect on the sample correlation. A practical management science application of the statistical procedures developed in this study is explored in the area of parametric cost estimation. (Author)

In the context of multiple linear regression, when a subset of k-out-of-p predictor variables is to be selected for the purpose of predicting the response at some known point in the predictor variables' space, the width of the resulting prediction interval gives an indication of the precision with which the response is predicted and, thus, it may provide a suitable selection criterion. A review of commonly used selection criteria is given, with special emphasis on those which deal with the problem of prediction. The Mahalanobis distance is one of the quantities affecting the width of the prediction interval, and it is studied in some detail. The effects of adding a new variable to a model are investigated and a monotonicity theorem is derived. The influence of an observation on the width of the prediction interval, as measured by the effected change when that observation is set aside, is also investigated and an equivalence beteween observation deletions and variable augmentation is shown.

This report discusses the concept, advantages and disadvantages of parametric cost estimation and points to the need for explicitly incorporating degree analogy between the system whose cost is to be predicted and the historical data base. Mahalanobis distance is proposed as a measure of analogy and its properties studied. A new variable selection criteria is proposed which incorporates both fit and analogy. An example is given to illustrate its use. (Author)

A scenario is described which involves ill-defined elements of conflict and cooperation: the acquisition of military systems by the Department of Defense (DoD) from large corporations. Current practices lead to situations in which DoD must deal with a sole source and thus forego savings which might be realized from competition between contractors. In order to deal with this situation, the Armed Services Procurement Regulations (ASPR) prescribes procedures which must be employed in the analysis and negotiation of sole source price proposals. These time-consuming procedures generate enormous proposal backlogs for government price analysts who, because of time pressure, may not be able to do a sufficiently thorough and accurate analysis upon which to base their negotiation position. This analysis paralysis also causes payment delays which, in turn, force contractors to borrow working capital and suffer capital costs. It is clearly in the best interests of all parties to expedite the processing of these proposals. This has been accomplished by developing statistical sampling and estimation techniques which, unlike some classical procedures, are not vulnerable to exploitation through the use of clever padding strategies.

Mahalanobis distance is suggested as a measure of analogy between a proposed system whose cost is to be predicted and a historical data base. It's role in the prediction problem and implications for selecting predictor variables is pointed out. (Author)

This paper concerns the distribution of the sum of k largest observations in a sample of m observations from a gamma distribution with n degrees of freedom. The density and cdf of the distribution are given as a sum of gamma density functions. If n is integer valued then the sum consists of a finite number terms. The distribution of the sum arises in a problem of selecting variables in a multiple regression analysis. (Author)

The distribution of the sum of K largest order statistics in a sample from the gamma distribution with n degrees of freedom is expressed as a linear function of gamma density functions if n is integer valued and is approximated by a probability distribution of the same form if n is not integer valued. (Author)

This report is a description of TSA, an interactive package of computer programs designed to implement Box-Jenkins type of analysis of univariate time series. In addition to the usual features found in many similar packages, TSA offers training modules which facilitate learning the 'art' of successful time series analysis. (Author)

For the problem of selecting the best population among several populations the intuitively natural selection rule is known to be optimal for a large class of loss functions and a family of distributions with monotone likelihood ratio property. In this paper a certain class of selection rules is considered and an optimal rule within the class is derived for an extended family of distributions.

The Basket Method of sampling, a tool designed to achieve statistically balanced samples, is described in intuitive terms. Special reference is made to applications in price analysis where experience has demonstrated the practicality of the technique. The intent is to provide an overview of what the system is intended to do and how it does it in order to assist price analysts and negotiators expedite proposal processing while maintaining acceptable levels of risk. Guidelines and examples are given for implementing a statistical pricing program tailored to local conditions. Underlying theory and documented computer codes are provided separately in Part I and Part III, respectively. (Author)

In this third of a series of three documents describing the basket method of sampling, source computer codes are provided for the convenience of users who may want to modify the programs to meet special local needs. (Author)

Efficient sequential stopping rules are developed which assure fixed length lower confidence limits on lot reliability when inspection is of the rectifying type. Several models of production processes are defined and analyzed: A nonparametric model; parametric models; parametric models with prior distributions. (Author)

A stopping rule for the process of sequential sampling inspection of finite lots is obtained which, with confidence 1-alpha, guarantees a prespecified upper confidence limit on the number of defective items in the uninspected portion of the lot. Usual assumptions of 'control' are not invoked here and, in that sense, the plans could be called nonparametric. Discussions of optimality and asymptotic results are included. (Author)

A minimax admissible decision rule is found for unscrambling the components of a random vector when a zero-one loss function is used. A loss is incurred if the estimated ordering of the components differs from the actual ordering. The rule is shown to be best invariant under the permutation group. (Author)