Calibrations developing

The data in the validation set are used to challenge the calibration. We treat the validation samples as if they are unknowns. We use the calibration developed with the training set to predict (or estimate) the concentrations of the components in the validation samples. We then compare these predicted concentrations to the actual concentrations as determined by an independent referee method (these are also called the expected concentrations). In this way, we can assess the expected performance of the calibration on actual unknowns. To the extent that the validation samples are a good representation of all the unknown samples we will encounter, this validation step will provide a reliable estimate of the calibration s performance on the unknowns. But if we encounter unknowns that are significantly different from the validation samples, we are likely to be surprised by the actual performance of the calibration (and such surprises are seldom pleasant). 

Calibration transfer, n - a method of applying a multivariate calibration developed on one instrument to data measured on a different instrument, by mathematically modifying the calibration model or by a process of instmment standardization. 

What near-infrared spectral region will be targeted for the calibration development, for example, combination region, first combination/overtone, or second combination/overtone Each region will imply a different transmission path length, and hence cell or probe design. 

Design of experiments (DOE) tools are widely considered to be some of the most important tools in the development of calibration models for PAT [22]. In cases where snfflciently relevant calibration samples can be synthetically prepared, DOE can be nsed to specify an efficient and effective set of calibration mixture standards for calibration development. Even in those PAT problems where synthetic standards cannot be prepared, the concepts behind DOE are nseful for understanding the importance of covering the needed composition and instrnment response space for a given problem. 

The fact that not all outliers are erroneous leads to the following suggested practice of handling outliers during calibration development (1) detect, (2) assess and (3) remove if appropriate. In practice, however, there could be hundreds or thousands of calibration samples and x variables, thus rendering individual detection and assessment of all outliers a rather time-consuming process. However, time-consuming as it may be, outlier detection is one of the most important processes in model development. The tools described below enable one to accomplish this process in the most efficient and effective manner possible. 

To minimize calibration development costs, it would be ideal to produce a calibration using data obtained on only one of the analyzers (the master instrument), and simply transfer this calibration to all of the other analyzers (the slave instruments). 

Beyers et aV° in the Polymer Research Division of BASF-AG used in-line transflectance NIR to monitor methyl methacrylate (MMA) and iV,7V-dimethylacrylamide (DMAAm) monomers in a copolymerization reaction. The work in this paper is of interest as it illustrates an example of calibration development done off-line with a very limited number of prepared calibration samples. The value of the measurement is to control the end properties of the products resulting from the copolymerization reaction. The end properties are related to many parameters including the intramolecular chemical composition distribution (CCD). The... 

Atomic force microscopy [6, 7] is one of the most suitable methods for research carbon nanotubes. AFM allows to receive not only a relief of the studied sample, but also distribution of mechanical characteristics, electric, magnetic and other properties on its surface. With the help of AFM, controllable manipulation of individual CNTs and CNTs bundles became possible. In this paper we report our approach to manipulating SWCNTs bundles with lateral force microscopy. LFM gives possibility to study lateral forces that probe acts upon bundles. In spite of good visualization of LFM, its lack is absence of reliable techniques of quantitative interpretation of results. The new way of calibration developed ourselves has allowed to pass from qualitative estimations to quantitative investigations [8], The given calibration technique is much more exact, than others known till now [9, 10], and does not assume simplification. With the help of new technique we may study adhesion of bundles to substrate and adhesion of CNTs in bundle qualitatively in real time more easy way. This result will provide new possibilities for nanotube application. 

Because of the favorable results obtained in the laboratory scale study, an experiment was conducted to assess the ability to predict the potency of pilot scale rotogranulated beads using data from the experiments performed in the laboratory. The calibration developed for the 55% laboratory scale beads was used to predict the potency of a 30-kg pilot batch. Figure 3 shows a plot of predicted versus actual potency. Prediction error for this study, although acceptable, was slightly higher than in the laboratory study. This error may be attributed to differences in surface characteristics and density between laboratory and pilot scale beads. 

Project funding for this Investigation was provided by the United Soybean Board (USB). Dr. Nick Bajjalieh was the Principle Investigator with contributions to calibration development and data assembly, analysis and release made by Caltest, LLC, Ballston Lake, NY. Source Bajjalieh, 2006. 

The USP chapter on Raman specifies that if possible the laser output should be monitored with a power meter from a reputable supplier. The variation of the laser power should be less than 25% and the laser power should be specified during the calibration development process. High laser power values can damage sensitive samples and low laser power values can yield Raman that are very noisy and are of low quality. 

---