build_krr_model: To build krr model

In conmolfields: Continuous molecular fields

Description

To build krr model

Usage

build_krr_model(K_train, y_train, gamma, set_b_0 = FALSE)

Arguments

K_train
y_train
gamma
set_b_0

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (K_train, y_train, gamma, set_b_0 = FALSE) 
{
    ntrain <- length(y_train)
    if (set_b_0) {
        if (rcond(K_train) < 1e-15) 
            return(NULL)
        a <- solve(K_train + gamma * diag(ntrain), y_train)
        b <- 0
    }
    else {
        K1 <- K_train + gamma * diag(ntrain)
        K2 <- cbind(K1, matrix(1, nrow = ntrain))
        K3 <- rbind(K2, matrix(1, ncol = ntrain + 1))
        K3[ntrain + 1, ntrain + 1] <- 0
        if (rcond(K3) < 1e-15) 
            return(NULL)
        y0 <- c(y_train, 0)
        ab <- solve(K3, y0)
        a <- ab[1:ntrain]
        b <- ab[ntrain + 1]
    }
    list(a = a, b = b)
  }

conmolfields documentation built on May 2, 2019, 4:18 p.m.