data-raw/05d_combine_models.md

In bennysalo/predict-recidivism: Analyses for prediction of recidivism in Finnish data using supervised learning

title: "Combine the best models of each type" author: "Benny Salo" date: "2019-02-14" output: github_document

Load the best models for each algorithm type. That is, the last updated models.

devtools::wd()
trained_mods_glm    <- readRDS("not_public/trained_mods_glm.rds")
trained_mods_glmnet <- readRDS("not_public/trained_mods_glmnet_3_LL.rds")
trained_mods_rf     <- readRDS("not_public/trained_mods_rf_3.rds")

Combine all models.

trained_mods_all <- c(trained_mods_glm, trained_mods_glmnet, trained_mods_rf)

Save the combined models.

devtools::wd()
saveRDS(trained_mods_all, "not_public/trained_mods_all.rds")

devtools::wd()
trained_mods_glm1000    <- readRDS("not_public/trained_mods_glm_4.rds")
trained_mods_glmnet1000 <- readRDS("not_public/trained_mods_glmnet_4.rds")
trained_mods_rf1000     <- readRDS("not_public/trained_mods_rf_4.rds")

Combine all models.

trained_mods_all1000 <- c(trained_mods_glm1000, 
                          trained_mods_glmnet1000, 
                          trained_mods_rf1000)

Save the combined models.

devtools::wd()
saveRDS(trained_mods_all1000, "not_public/trained_mods_all1000.rds")

bennysalo/predict-recidivism documentation built on May 29, 2019, 10:34 a.m.