Statistics of non-linear

A major disadvantage of the Solver is that it does not provide the standard deviations of the coefficients. This problem is addressed later in this chapter (see "Statistics of Non-linear Regression"). 

Another consideration when using the approach is the assumption that stress and strength are statistically independent however, in practical applications it is to be expected that this is usually the case (Disney et al., 1968). The random variables in the design are assumed to be independent, linear and near-Normal to be used effectively in the variance equation. A high correlation of the random variables in some way, or the use of non-Normal distributions in the stress governing function are often sources of non-linearity and transformations methods should be considered. 

Although the analogy is not perfect, this parameter can be thought of as a temperature in statistical physics or as the degree of non-linearity in a dynamical system. 

An extensive introduction into robust statistical methods is given in Ref. 134 a discussion of non-linear robust regression is found in Ref. 135. An example is worked in Section 3.4. 

Linearity should always be assessed initially by visual inspection of the plotted data and then by statistical evaluation. The linearity of the instrument response needs to be established because without this information it is difficult to attribute causes of non-linearity. The supporting statistical measures include correlation coefficient (r, r2, etc), residual plot, residual standard deviation and significance... 

However, a question arises - could similar approach be applied to chemical reactions At the first stage the general principles of the system s description in terms of the fundamental kinetic equation should be formulated, which incorporates not only macroscopic variables - particle densities, but also their fluctuational characteristics - the correlation functions. A simplified treatment of the fluctuation spectrum, done at the second stage and restricted to the joint correlation functions, leads to the closed set of non-linear integro-differential equations for the order parameter n and the set of joint functions x(r, t). To a full extent such an approach has been realized for the first time by the authors of this book starting from [28], Following an analogy with the gas-liquid systems, we would like to stress that treatment of chemical reactions do not copy that for the condensed state in statistics. The basic equations of these two theories differ considerably in their form and particular techniques used for simplified treatment of the fluctuation spectrum as a rule could not be transferred from one theory to another. 

For example, the standard synergetic approach [52-54] denies the possibility of any self-organization in a system with with two intermediate products if only the mono- and bimolecular reaction stages occur [49] it is known as the Hanusse, Tyson and Light theorem. We will question this conclusion, which in fact comes from the qualitative theory of non-linear differential equations where coefficients (reaction rates) are considered as constant values and show that these simplest reactions turn out to be complex enough to serve as a basic models for future studies of non-equilibrium processes, similar to the famous Ising model in statistical physics. Different kinds of auto-wave processes in the Lotka and Lotka-Volterra models which serve as the two simplest examples of chemical reactions will be analyzed in detail. We demonstrate the universal character of cooperative phenomena in the bimolecular reactions under study and show that it is reaction itself which produces all these effects. 

The thorough treatment of the experimental data does allow one to obtain reliable values of the reactivity ratios. The results of such a treatment are presented in Table 6.3 for some concrete system let us form a notion about an accuracy of the reactivity ratios estimations. The detailed analysis of such a significant problem in the case of the well-studied copolymerization of styrene with methyl methacrylate is reported in Ref. [227]. Important results on the comparison of the precision of rj, r2 estimates by means of different methods are presented by O Driscoll et al. [228]. Such a comparison of six well-known linear least-squares procedures [215-218,222,223] with the statistically correct non-linear least-squares method leads to the conclusion that some of them [216, 217, 222] can provide rather precise rls r2 estimates when the experiment is properly planned. 

Effect of temperature on the stability of foam bilayers. The effect of temperature on the rupture of foam bilayers has also been studied [414] with the help of microscopic NaDoS NB films with a radius of 250 pm. The dependence of the bilayer mean lifetime ton the surfactant concentration C in the presence of 0.5 mol dm 3 electrolyte (NaCI) at 10, 22 and 30°C has been obtained, the temperature being kept constant within 0.05°C. As in the above mentioned case, the NB foam films formed via black spots and the measurements were carried out after a sufficiently long time in order to allow equilibration of the system. At each of the NaDoS concentrations used and at the corresponding temperature, x was determined statistically and the comparison of the experimental with the theoretical x Q dependences was done by means of non-linear optimisation of the constants A, B and Ce. 

The Role of Non-linear Polarizability. We have just proved that the anisotropic molar saturation constants are non-zero even in liquids consisting of atoms and non-polar molecules. Beside the previous statistical-fluctuational processes, simple liquids reveal mechanisms that are due to the non-linear polarizabilities of atoms and such molecules. This is so because, in general, in addition to linear electric polarizability in the form (288), if a suffidently strong electric field is applied, we have to deal with the non-linear polarizability ... 

Booth" calculated the deviation of As from quadratidty by OnsagCT s model. Calculations of non-linear As-variations of higher order have also been poformed by Kielich in the Kirkwood-Frohlich semi-macroscopic approach taking into consideration statistical molecular correlations. Results such as these can be derived with the non-linear polarization (282). This treatment, however, is not directly applicable to the description of complete electric saturation, and we shall not develop it furtho- here. It appears preferable, for simplicity, to proceed within the framework of classical Langevin-Debye theory, which yidds results wdl adapted to numerical computations. ... 

To show the influence of various microscopic and structural factors on linear and non-linear effects in dense dielectrics, it is convenient to apply first a semi-macroscopic treatment of the theory, and then to proceed to its molecular-statistical interpretation, assuming appropriate microscopic models. The semi-macroscopic method was initially applied by Kirkwood and modified by Frohlich in the theory of linear dielectrics, and has beat successfully used in theories of non-linear tUelectrics. "... 

The determination of the concentration of the components (tT ) at the collocation points can be obtained by solving the set of non linear algebraic equations [5(N + 1)] resulting from insertion of equations (B.41-B.45) into equation (B.40) and excluding the surface where the concentrations are known. These equations are solved numerically by an IMSL (International Mathematical and Statistical Library) subroutine called ZSPOW based on a variation of Newton s method which uses a finite difference approximation to the Jacobian and takes precautions to avoid large step sizes or increasing residuals. 

The earliest research in a field builds on past work and this always makes it difficult to ascribe priority to important discoveries that lead to new directions or paradigms for future research. Certainly, there were many early investigations that used computer simulation of Newton s equations of motion to tackle important open problems. For instance, Hirschfelder et al. [2] studied the dynamics of the gas phase H + H2 reaction on a model potential surface to determine the reaction rate. In later years this investigation spawned the field of gas phase molecular dynamics. The paper by Fermi et al. [3] on the simulation of the dynamics of a model one-dimensional solid was influential in the field of non-linear dynamics. Neither these papers nor the body of work they stimulated had an immediate important impact on statistical mechanics [4]. 

The statistical mechanical treatment is carried out on the basis of Barker s self-consistent field theory. This treatment accounts for local order and yields a set of non linear equations, which are solved numerically by iterations. The most interesting point of this model is that it does not use adjustable parameters. However, the predictions are rather poor. 

When the EP comprises linear computations (linear in the observations) such as simple differences, y - B, or linear least squares or linear multivariate computations, initial normality (of the observations y) is preserved for the estimated quantities. Non-linear computations, such as arise commonly in iterative model selection and peak search routines, produce estimated parameters having non-normal distributions (59). Caution is in order, in those cases, in applying "normal" values of test statistics to calculate 1 and Cl s. (Other factors to consider are the extent of non-linearity, the level of confidence or significance [1-a], and the robustness of the statistic in question.)... 

Transient effects in the kinetics of oriented nucleation are considered for melt processing in a wide range of deformation rates using a theory of non-linear chain statistics with transient effects. Inverse Langevin elastic free energy of a polymer chain in a Pade approximation, averaged with transient distribution of the chain end-to-end vectors, as well as Peterlin s approximation for the modulus of nonlinear elasticity are used. The effects of transient orientation distribution of the chain segments is also considered. 

Non-linear models may be fitted to data sets by the inclusion of functions of physicochemical parameters in a linear regression model—for example, an equation in n and as shown in Fig. 6.5—or by the use of non-linear fitting methods. The latter topic is outside the scope of this book but is well covered in many statistical texts (e.g. Draper and Smith 1981). Construction of linear regression models containing non-linear terms is most often prompted when the data is clearly not well fitted by a linear model, e.g. Fig. 6.4e, but where regularity in the data suggests that some other model will fit. A very common example in the field of quantitative structure-activity relationship (QSAR) involves non-linear relationships with hydrophobic descriptors such as log P or n. Non-linear dependency of biological properties on these parameters became apparent early in the... 

Drenick, R. F., Park, C. B. (1975). Comments on worst inputs and a bound on the highest peak statistics of a class of non-linear systems. Journal of Sound and Vibration, 47,129-131. doi 10.1016/ S0022-460X(75)80203-3... 

M. Longuet-Higgins, The effect of non-linearities on statistical distirbutions in the theory of sea waves, J. Fluid Mechanics 17, 459-480 (1963). 

In the course of non-linear regression itself and the statistical design of experiments, one is very often confronted with severe numerical problems in optimization (and integration) caused by the shape of the response surface of the objective function /30/. 

---