Error in prediction

The largest errors in predicted compositions occur for the systems acetic acid-formic acid-water and acetone-acetonitrile-water where experimental uncertainties are significantly greater than those for other systems. 

Ideally, the results should be validated somehow. One of the best methods for doing this is to make predictions for compounds known to be active that were not included in the training set. It is also desirable to eliminate compounds that are statistical outliers in the training set. Unfortunately, some studies, such as drug activity prediction, may not have enough known active compounds to make this step feasible. In this case, the estimated error in prediction should be increased accordingly. 

The method is applicable at reduced temperatures above 0.30 or the freezing point, whichever is higher, and below the critical point. The method is most reliable when 0.5 prediction average 3.5 percent when experimental critical properties are known. Errors are higher for predic ted criticals. The method is useful when solved iteratively with Eq. (2-23) to predict the acentric factor. 

The Standard Error of Prediction (SEP) is supposed to refer uniquely to those situations when a calibration is generated with one data set and evaluated for its predictive performance with an independent data set. Unfortunately, there are times when the term SEP is wrongly applied to the errors in predicting y variables of the same data set which was used to generate the calibration. Thus, when we encounter the term SEP, it is important to examine the context in order to verify that the term is being used correctly. SEP is simply the square root of the Variance of Prediction, s2. The RMSEP (see below) is sometimes wrongly called the SEP. Fortunately, the difference between the two is usually negligible. 

Dus wall, A. A., Experimental Errors in Predictions of Thermal Hazards, Ibid., p. 55. 

The temperature distribution in the flow direction for a fixed flow rate differs for different devices. This suggests that the heat transfer mechanism in these devices is not identical. The non-uniform (of about 20%) heat flux leads to conditions at which the wall temperature increases sharply. Idealizing the heat flux as uniform can result in a significant error in prediction of the temperature distribution. 

Depending on the particular design of inlet and outlet manifolds, the difference between the flow rates into some parallel micro-channels may be up to 20%. Idealizing the flow rate as uniform can result in significant error in prediction of the temperature distribution of a heated electronic device. 

In the introduction we asserted that it was important to use the correct partition coefficients when interpreting U-series data. Both the ratio of daughter and parent partition coefficients and their absolute values are important. Small errors in the ratio can propagate to quite large errors in predictions of activity ratios even when the source material is assumed to have a parent-daughter ratio of unity (i.e., in radioactive... 

ACD/Labs have an extensive database which uses this approach. This approach works well except for anisotropic groups. Unlike carbon prediction this can have a massive effect on the chemical shift values and so can give rise to big errors in prediction, even for structural fragments that are well represented in... 

Component 1 2 mol% Range Component 2 Temperature Range, °C Pressure Range, Bars Absolute Avg Error in Predicted Water Content SRK PFGC ... 

Overall, this study indicated that generic simulation of pharmacokinetics at the lead optimization stage could be useful to predict differences in pharmacokinetic parameters of threefold or more based upon minimal measured input data. Fine discrimination of pharmacokinetics (less than twofold) should not be expected due to the uncertainty in the input data at the early stages. It is also apparent that verification of simulations with in vivo data for a few compounds of each new compound class was required to allow an assessment of the error in prediction and to identify invalid model assumptions. 

Based on errors in predicted A concentration. See Figure 5.71.) Components model ... 

The conclusion is that the model appears to be acceptable. This graph also provides information about how well the method will predict future samples. It is expected that the errors in prediction for component B will be 0.06. This conclu.sion is only possible because the validation set contains many samples that adequately span the calibration space (see Habit 1). A conclusion about the prediction errors for component A will be evaluated after resolving the issue with the unusual sample. 

The statistic used to quantift the error in prediction is the root mean square error of prediction (RMSEP) ... 

---