Trait linfa::metrics::SingleTargetRegression [−][src]
pub trait SingleTargetRegression<F: Float, T: AsTargets<Elem = F>>: AsTargets<Elem = F> {
fn max_error(&self, compare_to: &T) -> Result<F> { ... }
fn mean_absolute_error(&self, compare_to: &T) -> Result<F> { ... }
fn mean_squared_error(&self, compare_to: &T) -> Result<F> { ... }
fn mean_squared_log_error(&self, compare_to: &T) -> Result<F> { ... }
fn median_absolute_error(&self, compare_to: &T) -> Result<F> { ... }
fn r2(&self, compare_to: &T) -> Result<F> { ... }
fn explained_variance(&self, compare_to: &T) -> Result<F> { ... }
}
Expand description
Regression metrices trait for single targets.
It is possible to compute the listed mectrics between:
- One-dimensional array - One-dimensional array
- One-dimensional array - bi-dimensional array
- One-dimensional array - dataset
In the last two cases, if the second item does not represent a single target, the result will be an error.
To compare bi-dimensional arrays use MultiTargetRegression
Provided methods
fn mean_absolute_error(&self, compare_to: &T) -> Result<F>
fn mean_absolute_error(&self, compare_to: &T) -> Result<F>
Mean error between two continuous variables
fn mean_squared_error(&self, compare_to: &T) -> Result<F>
fn mean_squared_error(&self, compare_to: &T) -> Result<F>
Mean squared error between two continuous variables
fn mean_squared_log_error(&self, compare_to: &T) -> Result<F>
fn mean_squared_log_error(&self, compare_to: &T) -> Result<F>
Mean squared log error between two continuous variables
fn median_absolute_error(&self, compare_to: &T) -> Result<F>
fn median_absolute_error(&self, compare_to: &T) -> Result<F>
Median absolute error between two continuous variables
R squared coefficient, is the proportion of the variance in the dependent variable that is predictable from the independent variable
fn explained_variance(&self, compare_to: &T) -> Result<F>
fn explained_variance(&self, compare_to: &T) -> Result<F>
Same as R-Squared but with biased variance