max_dist_cost: Cost functions for automatic brokenstick models.

Description Usage Arguments Details Value See Also

Description

Given a model and data the function returns a single value which is used to determine if the number of points is optimized: the cost function value being minimized by the optimal set of parameters.

Usage

1
2
3
4
5
6
7
max_dist_cost(object)

dist_per_pt_cost(object)

rss_cost(object)

dzi_cost(object)

Arguments

object

Object of class inheriting from "bsm".

Details

max_dist_cost In this function the statistic is the maximun distance between an observation and its fitted value. Hence the function constantly decrease with increasing number of point in the model and a threshold must be provided along with this function to avoid infinite looping. Yet, the pros of this cost function is a threshold that is simple to determine and to interpret.

dist_per_pt_cost In this function the statistic is the average distance between observations and fitted values divided by the number of points used by the model. This function has does rech a minimun value but generally for high numbers of points.

rss_cost In this function the statistic is the sum of squared residuals.

dzi_cost In this function the statistic is the Dive Zone Index. See details in dive_zone_index

Value

A statistic to minimize.

See Also

optBrokenstick


SESman/rbl documentation built on May 9, 2019, 11:10 a.m.