compute_log_lik: Compute Negative Log Likelihood

In JohnNay/eat: Empirical Agent Training Software for Data-Driven Modeling

Description

Bernoulli likelihood: x = 0, 1; f(x,theta) = (theta^x)*(1-theta)^(1-x) Bernoulli log-likelihood: $$ ln(L) = sum_i=1^I sum_t=1^T D_i^C(t) * ln(P_i^C (t)) + (1 - D_i^C(t)) * ln(1 - P_i^C (t)) $$

Usage

compute_log_lik(prediction, actual)

Arguments

prediction	Numeric vector same length as actual
actual	Numeric vector same length as prediction

Value

Numeric vector length one

Examples

compute_log_lik(runif(10), sample(c(1,0), 10, replace=TRUE))

JohnNay/eat documentation built on May 7, 2019, noon