Formula: LL+ = -N+ log(1 - exp(-mu)) + log(mu) sum_i Y_i+ - N+ mu
sum_i log(Y_i+!) (eq. 15.2), the zero-truncated Poisson log-likelihood
used as the splitting criterion in the truncated part of the forest. With
x the best split is searched, taking the one that maximizes
LL+(left) + LL+(right).
Usage
Msm325(y_positive, mu = NULL, x = NULL)
Arguments
- y_positive
Strictly positive counts in the node.
- mu
Rate; the zero-truncated MLE is used when NULL.
- x
Candidate splitting covariate; no split is searched when NULL.
Value
List with estimate, loglik, mu, split, method.
References
Montesinos Lopez, Montesinos Lopez & Crossa (2022),
Multivariate Statistical Machine Learning Methods for Genomic Prediction,
Springer, eq. (15.2) p.651 and p.652. DOI 10.1007/978-3-030-89010-0.
Examples
# positive counts in a node, split on a covariate
Msm325(y_positive = c(1, 2, 1, 3, 4, 2, 6, 5), x = c(0.1, 0.2, 0.15, 0.4, 0.8, 0.3, 0.9, 0.7))$estimate
#> [1] -14.90951
