spring-class: Class "spring"
In jchiquet/spring: Fit the SPRING model

Description Slots Methods See Also

Class of object returned by the spring function.

coef.direct: List of pxq Matrices (with class "dgCMatrix") of estimated direct effects in the model with respect to the original input. The number of element in the list corresponds the length of lambda1.
coef.regres: List of pxq Matrices (with class "dgCMatrix") of estimated regression coefficients (both direct and indirect effects) in the model with respect to the original input. The number of element in the list corresponds the length of lambda1.
covariance: List of qxq matrices of estimated covariance between the output. The number of element in the list corresponds the length of lambda1.
intercept: List of qx1 vectors of estimated intercept terms. The number of element in the list corresponds the length of lambda1.
lambda1: Vector (class "numeric") of l1 norm penalty levels for which the model has eventually been fitted.
lambda2: Scalar for the level of (structuring) l2 norm penalty applied.
df: Estimated degree of freedoms for the successive lambda1.
residuals: List of Matrices of residuals. Each entry of the list corresponds to a value of lambda1.
r.squared:: Vector (class "numeric") given the coefficient of determination as a function of lambda1.
normx: Vector (class "numeric") containing the square root of the sum of squares of each column of the design matrix.
monitoring: List (class "list") which contains evolution of the loglikelihood and of the objective function for each lambda1 and each iteration of the algorithm.
fitted: List of Matrices of fitted values. Each entry of the list corresponds to a value of lambda1.

This class comes with the usual predict(object, newx, ...), fitted(object, ...), residuals(object, ...), print(object, ...), show(object) and deviance(object, ...) generic (undocumented) methods.

The two methods loglik(object) and objective(object) are also defined to recover the log-likelihood and objective functions along the fit. They can be applied on a spring object or on a list of spring object. The corresponding output is an object with class criterion. See the corrdesponding documentation for defintiion and methods.

The method spr.model(object, lambda1, lambda2=NULL)) applies on a spring object or a list of spring objects to extract an single object with class model. See the corresponding documentation for definition and methods.

See also spring, criterion, model.

jchiquet/spring documentation built on May 18, 2019, 10:22 p.m.