Adopted cross-validation technique

Both training and testing process take only several minutes on the PC equipped with the 486 processor in the case when up to 200 proteins are used. The FORTRAN source code, files with Gaussian parameters and protein data bases used in this report are available via Internet (see Supplementary Material). 

Conformational preference for transmembrane a-helix is strongly dependent on sequence hydrophobic environment for most amino acid types 

Asp and Glu have the F factor (a statistical measure for the dependence of preference on hydrophobicity of sequence neighbors) less than 50. 

Statistical parameters derived for the linear approximation of the dependence of helix 

Amino acid type Slope b Standard error in slope S), F parameter (b/si,)  

---

To avoid models with chance correlation, a check with different validation procedures must be adopted, such as, for example, cross-validation, y-scrambling and QUICK rule. A general validation procedure [Wold, 1991] would be the deletion of some objects before the selection of the variables applying the variable selection procedure and then predicting the responses for excluded objects. The whole procedure, including variable selection, is then repeated a number of times, depending on the adopted specific validation technique. 

The use of SVM cannot solve all problems of noise in data processing, but it is possible to use SVM technique to improve noisy data processing in many ways. For example, it can provide some ways for outlier deleting By leave-one-out (LOO) cross-validation method, we can delete the data samples with large error in prediction, and make the improvement of data files. Besides, the adoption of e-insensitive loss function in support vector regression makes it more robust to noisy data sets. 

---