adding files associated with manuscript

.Rbuildignore

@@ -14,3 +14,4 @@ README.Rmd
 README_files
 .github
 cran-comments.md
+^inst

.gitignore

@@ -3,3 +3,4 @@
 .RData
 .Ruserdata
 *.DS_Store
+*.Rproj

DESCRIPTION

@@ -1,10 +1,10 @@
 Package: veesa
 Type: Package
 Title: VEESA Pipeline for Explainable Machine Learning with Functional Data
-Version: 0.1.1
+Version: 0.1.2
 Author: Katherine Goode, J. Derek Tucker
 Maintainer: Katherine Goode <[email protected]>
-Description: Contains functions that are helpful when applying the VEESA pipeline in R.
+Description: Contains functions for applying the VEESA pipeline in R.
 License: MIT + file LICENSE
 Encoding: UTF-8
 LazyData: true

NEWS.md

@@ -1,8 +1,13 @@
+# veesa 0.1.2
+
+- Added shifted peaks data
+- Added example code from manuscript
+
 # veesa 0.1.1
 
 - Adjusted title in plot_pc_directions
 - Updated code to new version of fdasrvf
 
 # veesa 0.1.0
 
-Initial version of package
+Initial version of package

R/data_shifted_peaks.R

@@ -0,0 +1,9 @@
+#' "Shifted Peaks" Simulated Dataset
+#'
+#' A simulated dataset generated for examples in the veesa pipeline manuscript. 
+#' For the code used to prepare this dataset, see
+#' https://github.com/sandialabs/veesa/inst/data-shifted-peaks.md.
+#'
+#' @format A data.frame.
+
+"shifted_peaks"

inst/generate-data/shifted-peaks.md

@@ -0,0 +1,120 @@
+Shifted Peaks Data
+================
+
+Load packages:
+
+``` r
+library(dplyr)
+library(ggplot2)
+library(purrr)
+```
+
+Specify parameters:
+
+``` r
+z1 = 1; z1_sd = 0.05; a1 = -3; a1_sd = 1
+z2 = 1.25; z2_sd = 0.05; a2 = 3; a2_sd = 1
+```
+
+Specify the times:
+
+``` r
+times = seq(-15, 15, length.out = 150)
+```
+
+Functions:
+
+``` r
+# Function for simulating data
+generate_data = function(id, group, t) {
+  if (group == 1) {
+    z = rnorm(1, z1, z1_sd)
+    a = rnorm(1, a1, a1_sd)
+  }
+  if (group == 2) { 
+    z = rnorm(1, z2, z2_sd)
+    a = rnorm(1, a2, a2_sd)
+  }
+  y = z * exp((-((t - a)^2)) / 2)
+
+  data.frame(id, group = as.character(group), index = 1:length(t), t, y)
+}
+
+# Function for generating true means
+true_mean = function(group, t) {
+  if (group == 1) {
+   z = z1
+   a = a1 
+  }
+  if (group == 2) {
+    z = z2
+    a = a2
+  }
+  y = z * exp((-((t - a)^2)) / 2)
+  data.frame(group = as.character(group), index = 1:length(t), t, y)
+}
+```
+
+Specify the sample sizes:
+
+``` r
+ntrain = 400
+ntest = 100
+n = ntrain + ntest
+```
+
+Simulate data:
+
+``` r
+set.seed(20211027)
+shifted_peaks <-
+  map2_df(
+    .x = 1:n,
+    .y = c(rep(1:2, each = n / 2)),
+    .f = generate_data,
+    t = times
+  )
+```
+
+Randomly select ids from the simulated data for the training data:
+
+``` r
+train_ids = sample(unique(shifted_peaks$id), ntrain, F)
+```
+
+Add training/testing variable:
+
+``` r
+shifted_peaks <- 
+  shifted_peaks |>
+  mutate(data = ifelse(id %in% train_ids, "Training", "Testing")) |>
+  select(data, id, group, everything()) |>
+  mutate(data = forcats::fct_relevel(data, "Training", "Testing")) |>
+  mutate(group = factor(group))
+```
+
+Plot of the data:
+
+``` r
+ggplot(shifted_peaks, aes(x = t, y = y, color = group, group = id)) +
+  geom_line(alpha = 0.35) +
+  facet_grid(. ~ data) +
+  theme_bw() +
+  scale_color_manual(values = wesanderson::wes_palettes$Royal1[2:1]) +
+  ylim(0, 1.4) + 
+  guides(color = guide_legend(override.aes = list(alpha = 1))) +
+  labs(
+    color = "Group",
+    title = "Simulated functions",
+    x = "Time",
+    y = "y"
+  )
+```
+
+![](shifted-peaks_files/figure-commonmark/unnamed-chunk-9-1.png)
+
+Save the simulated data:
+
+``` r
+save(shifted_peaks, file = "../../data/shifted_peaks.rda")
+```

inst/generate-data/shifted-peaks.qmd

@@ -0,0 +1,123 @@
+---
+title: "Shifted Peaks Data"
+format: gfm
+---
+
+Load packages:
+
+```{r}
+#| message: false
+library(dplyr)
+library(ggplot2)
+library(purrr)
+```
+
+Specify parameters: 
+
+```{r}
+z1 = 1; z1_sd = 0.05; a1 = -3; a1_sd = 1
+z2 = 1.25; z2_sd = 0.05; a2 = 3; a2_sd = 1
+```
+
+Specify the times:
+
+```{r}
+times = seq(-15, 15, length.out = 150)
+```
+
+Functions: 
+
+```{r}
+# Function for simulating data
+generate_data = function(id, group, t) {
+  if (group == 1) {
+    z = rnorm(1, z1, z1_sd)
+    a = rnorm(1, a1, a1_sd)
+  }
+  if (group == 2) { 
+    z = rnorm(1, z2, z2_sd)
+    a = rnorm(1, a2, a2_sd)
+  }
+  y = z * exp((-((t - a)^2)) / 2)
+  
+  data.frame(id, group = as.character(group), index = 1:length(t), t, y)
+}
+
+# Function for generating true means
+true_mean = function(group, t) {
+  if (group == 1) {
+   z = z1
+   a = a1 
+  }
+  if (group == 2) {
+    z = z2
+    a = a2
+  }
+  y = z * exp((-((t - a)^2)) / 2)
+  data.frame(group = as.character(group), index = 1:length(t), t, y)
+}
+```
+
+Specify the sample sizes: 
+
+```{r}
+ntrain = 400
+ntest = 100
+n = ntrain + ntest
+```
+
+Simulate data: 
+
+```{r}
+set.seed(20211027)
+shifted_peaks <-
+  map2_df(
+    .x = 1:n,
+    .y = c(rep(1:2, each = n / 2)),
+    .f = generate_data,
+    t = times
+  )
+```
+
+Randomly select ids from the simulated data for the training data: 
+
+```{r}
+train_ids = sample(unique(shifted_peaks$id), ntrain, F)
+```
+
+Add training/testing variable:
+
+```{r}
+shifted_peaks <- 
+  shifted_peaks |>
+  mutate(data = ifelse(id %in% train_ids, "Training", "Testing")) |>
+  select(data, id, group, everything()) |>
+  mutate(data = forcats::fct_relevel(data, "Training", "Testing")) |>
+  mutate(group = factor(group))
+```
+
+Plot of the data:
+
+```{r}
+#| fig-width: 16
+#| fig-height: 4
+ggplot(shifted_peaks, aes(x = t, y = y, color = group, group = id)) +
+  geom_line(alpha = 0.35) +
+  facet_grid(. ~ data) +
+  theme_bw() +
+  scale_color_manual(values = wesanderson::wes_palettes$Royal1[2:1]) +
+  ylim(0, 1.4) + 
+  guides(color = guide_legend(override.aes = list(alpha = 1))) +
+  labs(
+    color = "Group",
+    title = "Simulated functions",
+    x = "Time",
+    y = "y"
+  )
+```
+
+Save the simulated data:
+
+```{r}
+save(shifted_peaks, file = "../../data/shifted_peaks.rda")
+```