Regularise set of profiles — smoothProfiles • SixSigma

This function takes a set of profiles and regularise them by means of a SVM

smoothProfiles(
  profiles,
  x = 1:nrow(profiles),
  svm.c = NULL,
  svm.eps = NULL,
  svm.gamma = NULL,
  parsvm.unique = TRUE
)

Arguments

profiles: Matrix of y values, one column per profile
x: Vector of predictive variable values, common to all profiles
svm.c: SVM parameter (cost)
svm.eps: SVM parameter (epsilon)
svm.gamma: SVM parameter (gamma)
parsvm.unique: Same parameters for all profiles? (logical [TRUE])

Value

Regularized profiles

Note

The package e1071 is needed in order to be able to use this function. SVM Parameters can be vectors of the same lenght as number of profiles, or a single value for all of them

References

Cano, E.L. and Moguerza, J.M. and Prieto Corcoba, M. (2015) Quality Control with R. An ISO Standards Approach. Springer.

Author

Javier M. Moguerza and Emilio L. Cano

Examples

wby.smooth <- smoothProfiles(profiles = ss.data.wby,
    x = ss.data.wbx)
plotProfiles(profiles = wby.smooth,
    x = ss.data.wbx)