Commit ⋅ ModelOriented/DALEX

Showing 2 of 6 files from the diff.

97.50% 100.00% +0.20%

R/predict_diagnostics.R changed.

100.00% 100.00% ø

R/plot_predict_diagnostics.R changed.

Other files ignored by Codecov

DESCRIPTION has changed.

NEWS.md has changed.

vignettes/vignette_titanic.Rmd has changed.

tests/testthat/test_predict_diagnostics.R has changed.

@@ -53,32 +53,38 @@

Loading

predict_diagnostics <-  function(explainer, new_observation, variables = NULL, ..., nbins = 20, neighbors = 50, distance = gower::gower_dist) {
  test_explainer(explainer, has_data = TRUE, function_name = "predict_diagnostics")

  neighbours_id <- select_neighbours_id(new_observation, explainer$data, n = neighbors, distance = distance)

  if (nrow(explainer$data) <= neighbors) {
    warning("Value of neighbors has to be lower than number of rows in explainer$data. Setting neighbors to nrow(explainer$data)")
    neighbors <- nrow(explainer$data) - 1
  }

  neighbors_id <- select_neighbors_id(new_observation, explainer$data, n = neighbors, distance = distance)


  # if variables = NULL then histograms with distribution of residuals are compared against each other
  if (is.null(variables)) {
    residuals_all <- explainer$residual_function(explainer$model, explainer$data, explainer$y, explainer$predict_function)
    residuals_sel <- residuals_all[neighbours_id]
    residuals_other <- residuals_all[-neighbours_id]
    residuals_sel <- residuals_all[neighbors_id]
    residuals_other <- residuals_all[-neighbors_id]

    cut_points <- signif(pretty(residuals_other, nbins), 3)
    test.res <- ks.test(residuals_other, residuals_sel)

    df1 <- data.frame(as.data.frame(table(cut(residuals_sel, cut_points))/length(residuals_sel)), direction = "neighbours")
    df1 <- data.frame(as.data.frame(table(cut(residuals_sel, cut_points))/length(residuals_sel)), direction = "neighbors")
    df2 <- data.frame(as.data.frame(-table(cut(residuals_other, cut_points))/length(residuals_other)), direction = "all")

    res <- list(variables = variables,
                histogram_neighbours = df1,
                histogram_neighbors = df1,
                histogram_all = df2,
                test = test.res,
                cut_points = cut_points,
                neighbours_id = neighbours_id)
                neighbors_id = neighbors_id)
  } else {
    # if variables is not null then we need to plot either categorical or continouse fidelity plot
    cp_neighbors <- ingredients::ceteris_paribus(explainer,
                                                 new_observation = explainer$data[neighbours_id, ],
                                                 y = explainer$y[neighbours_id],
                                                 new_observation = explainer$data[neighbors_id, ],
                                                 y = explainer$y[neighbors_id],
                                                 variables = variables,
                                                 ...)
    cp_new_instance <- ingredients::ceteris_paribus(explainer,

@@ -88,7 +94,7 @@

Loading

    res <- list(variables = variables,
                cp_neighbors = cp_neighbors,
                cp_new_instance = cp_new_instance,
                neighbours_id = neighbours_id)
                neighbors_id = neighbors_id)
  }
  class(res) <- "predict_diagnostics"
  res

@@ -99,7 +105,7 @@

Loading

individual_diagnostics <- predict_diagnostics


select_neighbours_id <- function(observation, data, variables = NULL, distance = gower::gower_dist, n = 50, frac = NULL) {
select_neighbors_id <- function(observation, data, variables = NULL, distance = gower::gower_dist, n = 50, frac = NULL) {
  if (is.null(variables)) {
    variables <- intersect(colnames(observation),
                           colnames(data))

@@ -34,7 +34,7 @@

Loading

plot.predict_diagnostics <- function(x, ...) {
  # if variables are not specified then gow with histogram
  if (is.null(x$variables)) {
    df <- rbind(x$histogram_neighbours, x$histogram_all)
    df <- rbind(x$histogram_neighbors, x$histogram_all)
    p.value <- x$test$p.value
    statistic <- x$test$statistic
    cut_points <- x$cut_points

Files		•	•	•	Coverage
R	1,021	889	0	132	87.07%
Project Totals (31 files)	1,021	889	0	132	87.07%

View CI Build Download

comment: false

coverage:
  status:
    project:
      default:
        target: auto
        threshold: 1%
        informational: true
    patch:
      default:
        target: auto
        threshold: 1%
        informational: true

Read the docs

Interact now

The inner-most circle is the entire project, moving away from the center are folders then, finally, a single file. The size and color of each slice is representing the number of statements and the coverage, respectively.

SVG

The top section represents the entire project. Proceeding with folders and finally individual files. The size and color of each slice is representing the number of statements and the coverage, respectively.

SVG

Each block represents a single file in the project. The size and color of each block is represented by the number of statements and the coverage, respectively.

SVG

Learn more about the coverage sunburst graph

53	53		predict_diagnostics <- function(explainer, new_observation, variables = NULL, ..., nbins = 20, neighbors = 50, distance = gower::gower_dist) {
54	54		test_explainer(explainer, has_data = TRUE, function_name = "predict_diagnostics")
55	55
56		-	neighbours_id <- select_neighbours_id(new_observation, explainer$data, n = neighbors, distance = distance)
	56	+
	57	+	if (nrow(explainer$data) <= neighbors) {
	58	+	warning("Value of neighbors has to be lower than number of rows in explainer$data. Setting neighbors to nrow(explainer$data)")
	59	+	neighbors <- nrow(explainer$data) - 1
	60	+	}
	61	+
	62	+	neighbors_id <- select_neighbors_id(new_observation, explainer$data, n = neighbors, distance = distance)
57	63
58	64
59	65		# if variables = NULL then histograms with distribution of residuals are compared against each other
60	66		if (is.null(variables)) {
61	67		residuals_all <- explainer$residual_function(explainer$model, explainer$data, explainer$y, explainer$predict_function)
62		-	residuals_sel <- residuals_all[neighbours_id]
63		-	residuals_other <- residuals_all[-neighbours_id]
	68	+	residuals_sel <- residuals_all[neighbors_id]
	69	+	residuals_other <- residuals_all[-neighbors_id]
64	70
65	71		cut_points <- signif(pretty(residuals_other, nbins), 3)
66	72		test.res <- ks.test(residuals_other, residuals_sel)
67	73
68		-	df1 <- data.frame(as.data.frame(table(cut(residuals_sel, cut_points))/length(residuals_sel)), direction = "neighbours")
	74	+	df1 <- data.frame(as.data.frame(table(cut(residuals_sel, cut_points))/length(residuals_sel)), direction = "neighbors")
69	75		df2 <- data.frame(as.data.frame(-table(cut(residuals_other, cut_points))/length(residuals_other)), direction = "all")
70	76
71	77		res <- list(variables = variables,
72		-	histogram_neighbours = df1,
	78	+	histogram_neighbors = df1,
73	79		histogram_all = df2,
74	80		test = test.res,
75	81		cut_points = cut_points,
76		-	neighbours_id = neighbours_id)
	82	+	neighbors_id = neighbors_id)
77	83		} else {
78	84		# if variables is not null then we need to plot either categorical or continouse fidelity plot
79	85		cp_neighbors <- ingredients::ceteris_paribus(explainer,
80		-	new_observation = explainer$data[neighbours_id, ],
81		-	y = explainer$y[neighbours_id],
	86	+	new_observation = explainer$data[neighbors_id, ],
	87	+	y = explainer$y[neighbors_id],
82	88		variables = variables,
83	89		...)
84	90		cp_new_instance <- ingredients::ceteris_paribus(explainer,

Navigation	Overlay
`t` Navigate files	`h` Toggle hits
`y` Change url to tip of branch	`m` Toggle misses
`b / v` Jump to prev/next hit line	`p` Toggle partial
`z / x` Jump to prev/next missed or partial line	`1..9` Toggle flags
`shift + o` Open current page in GitHub	`a` Toggle all on
`/ or ?` Show keyboard shortcuts dialog	`c` Toggle context lines or commits

88	94		res <- list(variables = variables,
89	95		cp_neighbors = cp_neighbors,
90	96		cp_new_instance = cp_new_instance,
91		-	neighbours_id = neighbours_id)
	97	+	neighbors_id = neighbors_id)
92	98		}
93	99		class(res) <- "predict_diagnostics"
94	100		res

99	105		individual_diagnostics <- predict_diagnostics
100	106
101	107
102		-	select_neighbours_id <- function(observation, data, variables = NULL, distance = gower::gower_dist, n = 50, frac = NULL) {
	108	+	select_neighbors_id <- function(observation, data, variables = NULL, distance = gower::gower_dist, n = 50, frac = NULL) {
103	109		if (is.null(variables)) {
104	110		variables <- intersect(colnames(observation),
105	111		colnames(data))

34	34		plot.predict_diagnostics <- function(x, ...) {
35	35		# if variables are not specified then gow with histogram
36	36		if (is.null(x$variables)) {
37		-	df <- rbind(x$histogram_neighbours, x$histogram_all)
	37	+	df <- rbind(x$histogram_neighbors, x$histogram_all)
38	38		p.value <- x$test$p.value
39	39		statistic <- x$test$statistic
40	40		cut_points <- x$cut_points

1	comment: false
2
3	coverage:
4	status:
5	project:
6	default:
7	target: auto
8	threshold: 1%
9	informational: true
10	patch:
11	default:
12	target: auto
13	threshold: 1%
14	informational: true