Expert systems as a tool for statistics: A review

The current state of using expert system (ES) techniques for statistics is expounded through the difficulties of defining the notion of statistical expert system (SES) and through several well-known examples. The main characteristics of statistical expertise, and particularly the important concept of strategy are emphasized. The current state of SESs is seen in the context of the global trend of the evolution of ES, i.e. within the growth of second generation expert system (SGESs). The knowledge base of these ESs represent an attempt to represent not only experts' knowledge but also the way experts acquire their expertise, thanks to the existence of an underlying model. We show through many examples how far SESs are and/or will be able to be SGESs.     

Paradigms and paradoxes: patterns and estimation of the entropy of formation of aqueous hydrogen containing mono and polynuclear oxyanions

“Simple” energetic patterns, where simple means the use of parameters derived only from the stoichiometry of these species, are relatively rarely discussed in the literature. In addition, entropy studies have been dominated by derivation of the absolute quantity S° rather than the entropy of formation (TΔ_f S ^o). Relationships between the entropy of formation and different parameters such as negative value of the charge of the species, the number of oxygen atoms, the natural logarithm of the molecular weight, the total number of atoms and the number of central atoms that are gases were recently discussed by us for aqueous polynuclear oxyanions. As shown here hydrogen containing anions do not follow this pattern. In this study, new approaches for the estimation of the entropy of formation of aqueous hydrogen containing mono and polynuclear oxyanions are suggested, evaluated and recommended.     

Statistical assertion: A more powerful method for debugging scientific applications

Traditional debuggers are of limited value for modern scientific codes that manipulate large complex data structures. Current parallel machines make this even more complicated, because the data structure may be distributed across processors, making it difficult to view/interpret and validate its contents. Therefore, many applications’ developers resort to placing validation code directly in the source program. This paper discusses a novel debug-time assertion, called a “Statistical Assertion”, that allows using extracted statistics instead of raw data to reason about large data structures, therefore help locating coding defects. In this paper, we present the design and implementation of an ‘extendable’ statistical-framework which executes the assertion in parallel by exploiting the underlying parallel system. We illustrate the debugging technique with a molecular dynamics simulation. The performance is evaluated on a 20,000 processor Cray XE6 to show that it is useful for real-time debugging.     

Equidistribution of negative statistics and quotients of Coxeter groups of type B and D

We generalize some identities and q-identities previously known for the symmetric group to Coxeter groups of type B and D. The extended results include theorems of Foata and Schützenberger, Gessel, and Roselle on various distributions of statistics, like inversion number, major index, and descent number. In order to show our results we provide explicit characterizations of the systems of minimal coset representatives of Coxeter groups of type B and D.     

A critical look at the definition of multidimensional separations

Multidimensional (MD) separations, especially comprehensive two-dimensional (2D) separations such as comprehensive 2D LC (LC × LC), and comprehensive 2D GC (GC × GC), are potentially powerful separation techniques. It is important to have a clear definition of MD techniques to better understand the scope and boundaries of the subject. Widely accepted definitions of MD Separations have their roots in the definition proposed by Giddings. Giddings also added several comments that clarified the scope of his definition. However, some researchers extend Giddings’ definitions beyond their intended scope. Doing so disqualifies such comprehensive 2D techniques as LC × LC, GC × GC and 2D TLC from being considered as 2D techniques. In other instances, extended treatment of Giddings’ definition is used as a basis to justify design-parameters of comprehensive 2D separations despite the fact that these parameters lead to sub-optimal implementations. We believe that the shortcomings in the definition and its popular interpretations are serious enough to warrant attention, especially by those interested in designing optimal instrumentation for MD separations like comprehensive 2D GC. After discussion of the weaknesses in the currently used definitions, we propose to define n-dimensional analysis as one that generates n-dimensional displacement information. We believe that this definition captures the spirit of Giddings’ definition while avoiding the problems associated with its popular interpretations.     

Emulator with inverse trend in first-order and second-order polarisation mode dispersion

We design a polarisation mode dispersion (PMD) emulator through subdividing a polarisation maintaining fibre (PMF) of 22 m in length. The aim of this emulator design is to show that first-order and second-order PMD can be inversely proportional to each other. Furthermore, the emulator is also used to show that the magnitude of PMD is independent to whether its statistics approach theoretical distributions or not, of most importance is the degree of mode coupling. The same (as former) applies to its autocorrelation function (ACF). The PMD control mechanism for the emulator is not in real time.     

Large deviation properties for patterns

Deciding whether a given pattern is over- or under-represented according to a given background model is a key question in computational biology. Such a decision is usually made by computing some p-values reflecting the “exceptionality” of a pattern in a given sequence or set of sequences. In the simplest cases (short and simple patterns, simple background model, small number of sequences), an exact p-value can be computed with a tractable complexity. The realistic cases are in general too complicated to get such an exact p-value. Approximations are thus proposed (Gaussian, Poisson, Large deviation approximations). These approximations are applicable under some conditions: Gaussian approximations are valid in the central domain while Poisson and Large deviation approximations are valid for rare events. In the present paper, we prove a large deviation approximation to the double strands counting problem that refers to a counting of a given pattern in a set of sequences that arise from both strands of the genome. In that case, dependencies between a sequence and its reverse complement cannot be neglected. They are captured here for a Bernoulli model from general combinatorial properties of the pattern. A large deviation result is also provided for a set of small sequences.     

数理统计与农业科学

戴君惕．数理统计与农业科学．数理统计与管理，１９９７，１６（４），４９～５２．本文介绍了数理统计学与农业科学的渊源及农业试验统计学的主要研究内容和应用，并讨论了数理统计学在农业科研和生产中的应用前景     

Bankruptcy prediction in banks and firms via statistical and intelligent techniques – A review

This paper presents a comprehensive review of the work done, during the 1968–2005, in the application of statistical and intelligent techniques to solve the bankruptcy prediction problem faced by banks and firms. The review is categorized by taking the type of technique applied to solve this problem as an important dimension. Accordingly, the papers are grouped in the following families of techniques: (i) statistical techniques, (ii) neural networks, (iii) case-based reasoning, (iv) decision trees, (iv) operational research, (v) evolutionary approaches, (vi) rough set based techniques, (vii) other techniques subsuming fuzzy logic, support vector machine and isotonic separation and (viii) soft computing subsuming seamless hybridization of all the above-mentioned techniques. Of particular significance is that in each paper, the review highlights the source of data sets, financial ratios used, country of origin, time line of study and the comparative performance of techniques in terms of prediction accuracy wherever available. The review also lists some important directions for future research.