Validation of predictions

A major difficulty in testing the validity of predictions from the DR equation is that independent estimates of the relevant parameters—the total micropore volume and the pore size distribution—are so often lacking. However, Marsh and Rand compared the extrapolated value for from DR plots of CO2 on a series of activated carbons, with the micropore volume estimated by the pre-adsorption of nonane. They found that except in one case, the value from the DR plot was below, often much below, the nonane figure (Table 4.9). 

K. Faber and B.R. Kowalski, Propagation of measurement errors for the validation of predictions obtained by principal component regession and partial least squares. J. Chemom., 11 (1997) 181-238. 

Figure 1.26 Experimental validation of predicted DCL behavior with metal-lomacrocycles.

We can state, in conclusion, that optimization of the potential energy function parameters on experimental data of small model compounds has led to a parameter set that gives an overall improvement of the accuracy of postdiction and, by implication, of the validity of prediction. The improvement is most marked in the most flexible substance. 

Persoone, G. Janssen, C.R. Field validation of predictions based on laboratory toxicity tests. In Freshwater Field Tests for Hazard Assessment of Chemicals Hill, I.A., Helmbach, F., Leeuwangh, P., Matthiessen, P., Eds. CRC Press, Inc. Boca Raton, EL, 1994 379-397. 

Particle Number Concentration and Size Distribution. The development of aerosol science to its present state has been directly tied to the available instrumentation. The introduction of the Aitken condensation nuclei counter in the late 1800s marks the beginning of aerosol science by the ability to measure number concentrations (4). Theoretical descriptions of the change in the number concentration by coagulation quickly followed. Particle size distribution measurements became possible when the cascade impactor was developed, and its development allowed the validation of predictions that could not previously be tested. The cascade impactor was originally introduced by May (5, 6), and a wide variety of impactors have since been used. Operated at atmospheric pressure and with jets fabricated by conventional machining, most impactors can only classify particles larger... 

This provides us with the first major test of the validity of predicted ozone depletion figures — how well do the observed and predicted middle and lower stratospheric OH concentrations in the present atmosphere correlate The conclusion is clear. There are no dependable middle and lower stratospheric OH observations. Thus, the exceedingly... 

Cairns Jr., J. (1986) What is meant by validation of predictions based on laboratory toxicity tests Hydrobiologia, 137, 271-278. 

Cairns Jr., J. (1988) What constitutes field validation of predictions based on laboratory evidence In Aquatic Toxicology andHazardAssessment, Adams, W.J., Chapman, G.A. and Landis, W.G. (eds), ASTM STP 971, pp. 361-368. American Society for Testing and Materials, Philadelphia, PA. 

To enable these computational tools to be physically relevant, they will be integrated with experimental efforts focused on the provision of accurate physical descriptions of new materials systems to increase the fidelity of predictive models, and critical validation of predictions. The integrated set of computational tools will significantly reduce extensive testing schedules and inspection procedures, and will contribute not only to development and optimization of new and existing materials within acceptable costs and time frames but also to component life extension programs. 

IAEA, Fuel Performance and Hssion Product Behavior in Gas-Cooled Reactors - A Compilation Produced within the IAEA Coordinated research program on Validation of Predictive Methods for Fuel and Fission Product Behavior in Gas-Cooled Reactors, IAEA-TECDOC-978, International Atomic Energy Agency, erma (1997). 

Validation of Predictive Methods for Fuel and Fission Product Behaviour in GCRs,... 

The experience base for GCR fuel behaviour under accident conditions was reviewed at an IAEA Specialists Meeting in 1990 [Ref. 2], and a CRP on Validation of Predictive Methods for Fuel and Fission Product Behaviour in GCRs was initiated in 1993. Countries participating in this CRP include China, France, Japan, Poland, Germany, the USA and the Russian Federation. Within this CRP, participants are documenting the status of the experimental data base and predictive methods, cooperating in methods verification and validation and will identify and document the additional needs for methods development and experimental validation data. 

Typically, QS AR models work reasonably well within their domain of application, but the predictivity breaks down significantly outside this domain. In pharmaceutical industry, the chemical estate is permanently increased by synthesizing or purchasing novel compounds, which are not necessarily well predicted by static models. Especially, if compounds are excluded from further activities based on in silico models, it is important to assess the validity of predictions. 

It is important to assess how well the predicting function fits the observed values of the dependent variable. Useful statistics are RSS in the case of a regression and TCE for classification problems. The calculation of these quantities is a resubstitutton, since the values used to obtain the predicting function are the same as the values used to validate the function. A typical statistic for validation of predicting function / in a regression is the multiple correlation coefficient... 

Coordinated research programmes on heat transfer and decay heat removal, reactor physics and validation of predictive methods for fuel behaviour are being conducted under the auspices of the IAEA. Actinide waste and plutonium burning gives another possible application for HTGRs and worldwide attention is being given to this possibility. 

As an example of the general validity of prediction 1, Hildebrand and Scott list the following solubilities of gaseous CI2 in several dissimilar solvents at 0 °C and a partial pressure of 1.01 bar xb = 0.270 in heptane, xb = 0.288 in SiCU, and xb = 0.298 in CCI4. These values are similar to one another and close to the ideal value p- /= 0.273. 

This simple approach was used in the plumbosolvency control programme in Wales (Hayes et al., 2008) and enabled a zonal emission model to be calibrated that good validation of predicted RDT sample results was obtained from actual RDT sample results indicated that this simple methodology was adequate for estimating the percentage of houses with a lead pipe. However, to put this into perspective, over 11,000 results were available across 29 water supply systems, with an average of 383 RDT sample results per scheme. 

Validation of predictive methods for fuel and fission product behaviour in gas cooled... 

In most cases, the validation of predictions from accelerated ageing remains very difficult as real-time data from field trials remains very limited and there remains significant uncertainty on how to define the simulated age (i.e. artificially accelerated age) of multi-material assemblies. 

Criteria have been proposed to define the reliability of simulation results as well as to evaluate the validity of prediction methods. The estimation of rms deviations between the native and the simulated structure of a protein is not a sufficient criterium for such an evaluation. It provides a single number for the goodness of fit of many pieces of numerical information which describes the conformation. Different methods may be used to determine the success of simulation procedure they allow one to compare either the overall conformation, or local structures, or both. The method used by... 

---