add new tutorial sdd2.02a

Author: GuyliannEngels

inst/tutorials/02a_reg_multi/reg_lin_multi.Rmd

@@ -5,7 +5,7 @@ output:
   learnr::tutorial
 tutorial:
   id: "sdd2.02a"
-  version: 0.1.0
+  version: 1.0.0
 runtime: shiny_prerendered
 ---
 
@@ -152,8 +152,8 @@ quiz(
     allow_retry = TRUE, random_answer_order = TRUE
     ),
   question(text = "Quelle est la valeur de la pente ?",
-    answer(sprintf("%.2f", 0), correct = TRUE),
-    answer(sprintf("%.2f", lm_lin_result$estimate[1])),
+    answer(sprintf("%.2f", 0)),
+    answer(sprintf("%.2f", lm_lin_result$estimate[1]), correct = TRUE),
     answer(sprintf("%.2f", lm_lin_param$BIC[1])),
     answer(sprintf("%.2f", lm_lin_param$AIC[1])),
     answer(sprintf("%.2f", lm_lin_param$r.squared[1])),
@@ -176,39 +176,106 @@ quiz(
 # edition de l'exercice 
 set.seed(42)
 
+vec <- seq(from = 1, to = 15, by = .2) 
+x <- vec + rnorm(sd = 3, n = length(vec))
 
-a <- seq(from = 100, to = 500) 
-random <- rnorm(n = length(a), sd = 50)
-x <- a + random
-x0 <- x + rnorm(n = length(x))
-x1 <- 1/x 
-x2 <- log(x)
-x3 <- exp(x)
-x4 <- sin(x)
-x5 <- cos(x)
+reg_lin_rand <- function(x, a, b, random = 0.5){
+  y <- (a*x) + b + rnorm(sd=random, n = length(x))
+  y 
+}
+reg_mutli3_rand <- function(x1, a1, x2, a2, x3, a3, b, random = 5){
+  y <- a1*x1 + a2*x2 + a3*x3 + b + rnorm(sd=random, n = length(x1))
+  y
+}
+
+df2 <- tibble::tibble(
+  x = x,
+  x0 = reg_lin_rand(x = x, a = 1.26, b = 2, random = 3.5),
+  x1 = reg_lin_rand(x = x, a = 1.5, b = 1, random = 10),
+  y = reg_mutli3_rand(x1 = x,
+                x2 = x0, 
+                x3 = x1,
+                a1 = 0.2, a2 = 0.3, a3 = 1.1,
+                b = 2))
+
+
+lm_mult <- lm(data = df2, formula = y  ~ x + x1 )
+lm_mult_coef <- broom::tidy(lm_mult)
+lm_mult_param <- broom::glance(lm_mult)
+```
 
-df <- tibble::tibble(x, x0, x1, x2, x3, x4, x5)
+Réalisez une régression linéaire simple sur le jeu de données `df2` de la variable y en fonction de la variable `x` et `x1`
 
-plot(df)
-corrplot::corrplot(cor(df,
-  use = "pairwise.complete.obs"), method = "ellipse")
+```{r regmulti-prep}
+# edition de l'exercice 
+set.seed(42)
 
-corrplot::corrplot(cor(df,
-  use = "pairwise.complete.obs", method = "spearman"), method = "ellipse")
+vec <- seq(from = 1, to = 15, by = .2) 
+x <- vec + rnorm(sd = 3, n = length(vec))
 
-mod_poly2 <- function(x1, alpha1, alpha2, intercept, random_effect){
-  y <- intercept + (alpha1 * x1) + (alpha2 * (x1^2)) 
-  y + rnorm(n = length(x1), sd = random_effect)
+reg_lin_rand <- function(x, a, b, random = 0.5){
+  y <- (a*x) + b + rnorm(sd=random, n = length(x))
+  y 
+}
+reg_mutli3_rand <- function(x1, a1, x2, a2, x3, a3, b, random = 5){
+  y <- a1*x1 + a2*x2 + a3*x3 + b + rnorm(sd=random, n = length(x1))
+  y
 }
 
-df <- tibble(
-  x = x1,
-  y = mod_poly2(x1 = x1, alpha1 = 2, alpha2 = 2.5, intercept = 55, random_effect = 10)
-)
+df2 <- tibble::tibble(
+  x = x,
+  x0 = reg_lin_rand(x = x, a = 1.26, b = 2, random = 3.5),
+  x1 = reg_lin_rand(x = x, a = 1.5, b = 1, random = 10),
+  y = reg_mutli3_rand(x1 = x,
+                x2 = x0, 
+                x3 = x1,
+                a1 = 0.2, a2 = 0.3, a3 = 1.1,
+                b = 2))
+```
+
+```{r regmulti, exercise = TRUE, exercise.setup = "regmulti-prep"}
+# résumé des données 
+df2
+#
+```
 
-lm_poly <- lm(df, formula = y ~ x + I(x^2))
-lm_poly_coef <- broom::tidy(lm_poly)
-lm_poly_param <- broom::glance(lm_poly)
+```{r regmulti-hint-1}
+# Snippet
+summary(lm. <- lm(data = DF, FORMULA))
+```
+
+```{r regmulti-hint-2}
+summary(lm. <- lm(data = DF, formula =  Y  ~ VAR1 + VAR2))
+```
+
+```{r regmulti-solution}
+summary(lm. <- lm(data = df2, formula = y ~ x + x1))
+```
+
+```{r regmulti-check}
+#TODO
+```
+
+Suite à votre analyse répondez aux questions suivantes : 
+
+```{r qu_regmulti}
+quiz(
+  question(text = "Quelle est la valeur de l'ordonnée à l'origine ?",
+    answer(sprintf("%.2f", lm_mult_coef$estimate[1]), correct = TRUE),
+    answer(sprintf("%.2f", lm_mult_coef$estimate[2])),
+    answer(sprintf("%.2f", lm_mult_coef$p.value[1])),
+    answer(sprintf("%.2f", lm_mult_param$AIC[1])),
+    answer(sprintf("%.2f", lm_mult_param$r.squared[1])),
+    allow_retry = TRUE, random_answer_order = TRUE
+    ),
+  question(text = "Quelle est la fraction de la variance exprimée par la régression linéaire ?",
+    answer(sprintf("%.3f", lm_mult_param$adj.r.squared), correct = TRUE),
+    answer(sprintf("%.3f", lm_mult_param$r.squared)),
+    answer(sprintf("%.3f", lm_mult_param$df)),
+    answer(sprintf("%.3f", lm_mult_coef$estimate[1])),
+    allow_retry = TRUE, random_answer_order = TRUE
+    )
+)
 ```
 
 ## Conclusion