CombinePreds: Combine predictions from multiple models: for example combine...

In cities-lab/VETravelDemandMM: Simulate Multi-Modal Travel Demand for Households

Description

Combine predictions from multiple models: for example combine GreenSTEP zero dvmt glm model and dvmt lm model As an example (a clumsy way to square mpg):

Usage

CombinePreds(preds_ls, func = `*`, init = 1, ...)

Arguments

preds_ls	A list of prediction vectors for each modeling steps
func	function used to combine results from modeling steps
init	initial value, pass to the '.init' of 'reduce()'
...	additional argument passed to 'reduce'

Value

a list with the combined predictions

Examples

 library(dplyr)
 library(purrr)
 library(tidyr)
 mtcars %>%
   nest(-cyl) %>%
   crossing(tibble(x=c(1, 2))) %>%
   mutate(mpg=map(data, "mpg")) %>%
   group_by(cyl) %>%
   summarise(mpg1 = list(mpg[[1]]),
             mpg2 = CombinePreds(mpg)) %>%
   mutate(test=map2_lgl(mpg1, mpg2, ~all(.y==.x^2)))

cities-lab/VETravelDemandMM documentation built on Aug. 1, 2019, 4:43 p.m.