Data rejection

The EPA recommends that advisory recovery acceptance criteria of 70-130 percent be used until laboratory control limits become available (EPA, 1996a) however, many laboratories use arbitrary limits as data quality acceptance criteria. The arbitrary selection of acceptance criteria for accuracy and precision is a harmful practice on the part of the laboratory and the data user. Arbitrarily criteria that are too narrow lead to unnecessary data rejection criteria that are too wide damage data... 

One of the criteria used for data rejection is Chauvenet s criterion, stated as follows ... 

First set of data rejected from statistical analysis. ... 

Hypothetical projections of the solidus surfaces of the Sc-M-C (M=Zr, Hf, V, Nb, Ta) phase diagrams are proposed in a paper of Velikanova et al. (1989). A complete miscibility of the binary carbides ScCi- and MCi- (M=Zr, Hf, V, Nb, Ta) crystallizing in the NaCl type structure, and the absence of ternary compounds in each of these ternary systems is predicted. However, these data need to be confirmed experimentally. Recently Ilyenko et al. (1996) reported the existence of continuous solid solutions between ScCi x and MC] x (M=Zr, Hf), and accordingty these solid solutions are in equilibria with all the phases of the solidus surface. The same authors reported that the VCi x binary carbidedissolves at least 15at.% Sc while the V solubility in ScC x is at least 17at.%. These data reject the earlier prediction of a complete miscibility of the NaCl-t5q)e binary carbides in the system Sc-V-C. However, no figures of the phase equilibria are presented in the brief communication of Ilyenko et al. 

An additional advantage derived from plotting the residuals is that it can aid in detecting a bad data point. If one of the points noticeably deviates from the trend line, it is probably due to a mistake in sampling, analysis, or reporting. The best action would be to repeat the measurement. However, this is often impractical. The alternative is to reject the datum if its occurrence is so improbable that it would not reasonably be expected to occur in the given set of experiments. 

The secret to success has been to learn from data and from experiments. Chemists have done a series of experiments, have analyzed them, have looked for common features and for those that are different, have developed models that made it po.ssiblc to put these observation.s into a systematic ordering scheme, have made inferences and checked them with new experiments, have then confirmed, rejected, or relined their models, and so on. This process is called inductive learning (Figure 1 -1), a method chemists have employed from the veiy beginnings ol chcmistiy. 

As explained in Chapter 8, descriptors are used to represent a chemical structure and, thus, to provide a coding which allows electronic processing of chemical data. The example given here shows how a GA is used to Rnd an optimal set of descriptors for the task of classification using a Kohoncii neural network. The chromosomes of the GA are to be used as a means for selecting the descriptors they indicate which descriptors are used and which are rejected ... 

Next, an equation for a test statistic is written, and the test statistic s critical value is found from an appropriate table. This critical value defines the breakpoint between values of the test statistic for which the null hypothesis will be retained or rejected. The test statistic is calculated from the data, compared with the critical value, and the null hypothesis is either rejected or retained. Finally, the result of the significance test is used to answer the original question. 

The value of fexp is then compared with a critical value, f(a, v), which is determined by the chosen significance level, a, the degrees of freedom for the sample, V, and whether the significance test is one-tailed or two-tailed. For paired data, the degrees of freedom is - 1. If fexp is greater than f(a, v), then the null hypothesis is rejected and the alternative hypothesis is accepted. If fexp is less than or equal to f(a, v), then the null hypothesis is retained, and a significant difference has not been demonstrated at the stated significance level. This is known as the paired f-test. 

This experiment uses the change in the mass of a U.S. penny to create data sets with outliers. Students are given a sample of ten pennies, nine of which are from one population. The Q-test is used to verify that the outlier can be rejected. Glass data from each of the two populations of pennies are pooled and compared with results predicted for a normal distribution. 

Some data iEustrating the effect of pressure on the water and salt fluxes and the salt rejection of a good quaUty reverse osmosis membrane are shown ia Figure 34 (76). 

Subdivision O guidelines for residue chemistry data were originally pubHshed by the EPA in 1982. These have been supplemented to improve the rate of acceptance by EPA reviewers of the many reports submitted by registrants in support of tolerances for pesticides in foods. The residue chemistry studies most frequently rejected include metaboHsm in plants, food processing (qv) studies, and studies on storage stabHity of residues in field samples (57). AH tolerances (maximum residue levels) estabHshed under FIFRA are Hsted in 40 CFR under Sections 180 for individual pesticides in/on raw agricultural commodities, 180 for exemptions from tolerances, 185 for processed foods, and 186 for animal feeds. 

Using this simplified model, CP simulations can be performed easily as a function of solution and such operating variables as pressure, temperature, and flow rate, usiag software packages such as Mathcad. Solution of the CP equation (eq. 8) along with the solution—diffusion transport equations (eqs. 5 and 6) allow the prediction of CP, rejection, and permeate flux as a function of the Reynolds number, Ke. To faciUtate these calculations, the foUowiag data and correlations can be used (/) for mass-transfer correlation, the Sherwood number, Sb, is defined as Sh = 0.04 S c , where Sc is the Schmidt... 

The intrinsic rejection and maximum obtainable water flux of different membranes can be easily evaluated in a stirred batch system. A typical batch unit (42) is shown in Figure 5. A continuous system is needed for full-scale system design and to determine the effects of hydrodynamic variables and fouling in different module configurations. A typical laboratory/pilot-scale continuous unit using computer control and on-line data acquisition is shown in Figure 6. 

The infrared spectmm of caprolactam has been given (3). Melting point data for the caprolactam—water system, as shown in Eigute 1, ate indicative of successful purification of caprolactam by crystallization from aqueous solution such purification is very effective for separating and rejecting polar impurities. 

Risk and uncertainty associated with each venture should translate, ia theory, iato a minimum acceptable net return rate for that venture. Whereas this translation is often accompHshed implicitly by an experienced manager, any formal procedure suffers from the lack of an equation relating the NRR to risk, as well as the lack of suitable risk data. A weaker alternative is the selection of a minimum acceptable net return rate averaged for a class of proposed ventures. The needed database, from a collection of previous process ventures, consists of NPV, iavestment, venture life, inflation, process novelty, decision (acceptance or rejection), and result data. 

It is common that compressor manufacturers provide data for the ratio of the heat rejected at the condenser to the refrigeration capacity as shown in Fig. 11-89. The solid line represents data for the open compressors while the dotted hne represents the hermetic and accessible compresors. The difference between sohd and dotted line is due to all losses (mechanical and elec trical in the electrical motor). Condenser design is based on the value ... 

Cropley made general recommendations to develop kinetic models for compUcated rate expressions. His approach includes first formulating a hyperbolic non-linear model in dimensionless form by linear statistical methods. This way, essential terms are identified and others are rejected, to reduce the number of unknown parameters. Only toward the end when model is reduced to the essential parts is non-linear estimation of parameters involved. His ten steps are summarized below. Their basis is a set of rate data measured in a recycle reactor using a sixteen experiment fractional factorial experimental design at two levels in five variables, with additional three repeated centerpoints. To these are added two outlier... 

Summaries of the data resourees considered useful were prepared. Useful data was defined as information that was publicly availdWe, scientifically collected, had statistical merit, and could be used for CPQRAs. A list of rejected resources was retained to identify references for supplemental reading and to avoid review duplication when the anticipated second edition of this book is developed. In total, 72 resources were accepted, and over 200 references were rejected. 

The selection of data from the resources available required decisions about the acceptable quality of the data and applicability of the data to the CCPS Taxonomy. Data from a resource was rejected by SAIC and the CCPS Subcommittee when ... 

It should be noted that data were not rejected through consideration of upper or lower bounds. These limits for the input data included a variety of assumed and calculated limits using various levels of confidence. 

---