## ----include = FALSE----------------------------------------------------------
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 6,
  fig.height = 4
)
set.seed(1)

## -----------------------------------------------------------------------------
library(combss)

## -----------------------------------------------------------------------------
set.seed(102)
n <- 300; p <- 30
beta <- c(3, 2, 1.5, 1, 0.5, rep(0, p - 5))
x <- matrix(rnorm(n * p), n, p)
y <- as.numeric(x %*% beta + rnorm(n) * 0.5)

itr <- 1:200; iva <- 201:300
fit <- combss(x[itr, ], y[itr],
              x_val = x[iva, ], y_val = y[iva],
              family = "gaussian", q = 10)
fit

## -----------------------------------------------------------------------------
fit$subset_list[1:8]

## -----------------------------------------------------------------------------
plot(seq_along(fit$mse_path), fit$mse_path, type = "b",
     xlab = "k", ylab = "Validation MSE")
abline(v = fit$best_k, lty = 2)

## -----------------------------------------------------------------------------
ybin <- as.numeric(plogis(x %*% beta) > 0.5)
fit_bin <- combss(x[itr, ], ybin[itr],
                  x_val = x[iva, ], y_val = ybin[iva],
                  family = "binomial", q = 10)
fit_bin

## -----------------------------------------------------------------------------
ymulti <- cut(as.numeric(x %*% beta),
              breaks = c(-Inf, -1, 1, Inf),
              labels = c("a", "b", "c"))
fit_mn <- combss(x[itr, ], ymulti[itr],
                 x_val = x[iva, ], y_val = ymulti[iva],
                 family = "multinomial", q = 10)
fit_mn

## ----eval = FALSE-------------------------------------------------------------
# cv <- combss_cv(x, y, family = "gaussian", q = 6)
# cv$best_lambda

## -----------------------------------------------------------------------------
preds <- predict(fit, newx = x[iva, ],
                 x_train = x[itr, ], y_train = y[itr])
head(preds)