run-all-lasso-stand.R

# do analysis for different data sets and combinations
getwd()
setwd("Repos/MA-climate/")
source("code/R/helper-functions.R")
library(raster)
library(ggplot2)
library(glmnet)

save_to <- "test-lasso-stand"
sst_cv_path <- "data/processed/sst_cv.rds"
precip_cv_path <- "data/processed/precip_cv.rds"
standardize_features <- FALSE
standardize_response <- TRUE
sst_eval_path <- "data/processed/sst_eval.rds"
precip_eval_path <- "data/processed/precip_eval.rds"
center_response <- FALSE
diff_n <- FALSE

sst_cv <- readRDS("data/processed/sst_cv.rds")
precip_cv <- readRDS("data/processed/precip_cv.rds")

test_lasso <- cv_for_ts(sst = sst_cv, precip = precip_cv, nfold = 5, 
                        size_train = 60, size_test = 14,
                        save_folder = save_to,
                        model = "lasso", 
                        include_ts_vars=FALSE,
                        stand=FALSE, diff_features=FALSE, 
                        des_features=FALSE,
                        center_response=FALSE,
                        standardize_features = standardize_features)

err_bar_plot <- readRDS(paste0("results/CV-lasso/",save_to,"/err-mat-plots/err-bars-plot.rds"))
ll <- get_l_1se_high_low(err_bar_plot)

fit_eval_full_lasso(save_to, sst_cv_path, precip_cv_path,
                    sst_eval_path, precip_eval_path,
                    standardize_response = standardize_response,
                    center_response = center_response,
                    standardize_features = standardize_features,
                    diff_n = diff_n)