Skip to content

Commit 6a591bc

Browse files
JohnMountJohnMount
committed
rebuild examples
1 parent a98eeb7 commit 6a591bc

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

43 files changed

+721
-503
lines changed

‎Examples/Classification/Classification.Rmd‎

Lines changed: 12 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,7 @@ Load modules/packages.
1818
```{r}
1919
library(rqdatatable)
2020
library(vtreat)
21+
packageVersion('vtreat')
2122
suppressPackageStartupMessages(library(ggplot2))
2223
library(WVPlots)
2324
```
@@ -30,6 +31,8 @@ Generate example data.
3031
* Input `x2` is a pure noise variable with no relationship to the output
3132

3233
```{r}
34+
set.seed(2020)
35+
3336
make_data <- function(nrows) {
3437
d <- data.frame(x = 5*rnorm(nrows))
3538
d['y'] = sin(d['x']) + 0.1*rnorm(n = nrows)
@@ -97,9 +100,16 @@ score_frame <- transform$scoreFrame
97100
score_frame$recommended <- score_frame$varMoves & (score_frame$sig < 1/nrow(score_frame))
98101
```
99102

100-
Note that for the training data `d`: `transform_design$crossFrame` is **not** the same as `transform.prepare(d)`; the second call can lead to nested model bias in some situations, and is **not** recommended.
103+
Note that for the training data `d`: `transform_design$crossFrame` is **not** the same as `prepare(transform, d)`; the second call can lead to nested model bias in some situations, and is **not** recommended.
101104
For other, later data, not seen during transform design `transform.preprare(o)` is an appropriate step.
102105

106+
`vtreat` version `1.5.1` and newer issue a warning if you call the incorrect transform pattern on your original training data:
107+
108+
```{r}
109+
d_prepared_wrong <- prepare(transform, d)
110+
```
111+
112+
103113
Now examine the score frame, which gives information about each new variable, including its type, which original variable it is derived from, its (cross-validated) correlation with the outcome, and its (cross-validated) significance as a one-variable linear model for the outcome.
104114

105115
```{r}
@@ -204,6 +214,7 @@ dtest <- make_data(450)
204214
205215
# prepare the new data with vtreat
206216
dtest_prepared = prepare(transform, dtest)
217+
# dtest %.>% transform is an alias for prepare(transform, dtest)
207218
208219
# apply the model to the prepared data
209220
dtest_prepared['prediction'] = predict(

‎Examples/Classification/Classification.md‎

Lines changed: 124 additions & 86 deletions
Large diffs are not rendered by default.

‎Examples/Classification/Classification_FT.Rmd‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,7 @@ Load modules/packages.
1818
```{r}
1919
library(rqdatatable)
2020
library(vtreat)
21+
packageVersion('vtreat')
2122
suppressPackageStartupMessages(library(ggplot2))
2223
library(WVPlots)
2324
```
@@ -30,6 +31,8 @@ Generate example data.
3031
* Input `x2` is a pure noise variable with no relationship to the output
3132

3233
```{r}
34+
set.seed(2020)
35+
3336
make_data <- function(nrows) {
3437
d <- data.frame(x = 5*rnorm(nrows))
3538
d['y'] = sin(d['x']) + 0.1*rnorm(n = nrows)

‎Examples/Classification/Classification_FT.md‎

Lines changed: 51 additions & 43 deletions
Original file line numberDiff line numberDiff line change
@@ -28,6 +28,12 @@ library(rqdatatable)
2828

2929
``` r
3030
library(vtreat)
31+
packageVersion('vtreat')
32+
```
33+
34+
## [1] '1.5.1'
35+
36+
``` r
3137
suppressPackageStartupMessages(library(ggplot2))
3238
library(WVPlots)
3339
```
@@ -44,6 +50,8 @@ Generate example data.
4450
<!-- end list -->
4551

4652
``` r
53+
set.seed(2020)
54+
4755
make_data <- function(nrows) {
4856
d <- data.frame(x = 5*rnorm(nrows))
4957
d['y'] = sin(d['x']) + 0.1*rnorm(n = nrows)
@@ -62,14 +70,14 @@ d %.>%
6270
knitr::kable(.)
6371
```
6472

65-
| x | y | xc | x2 | yc |
66-
| ----------: | ----------: | :---------- | ----------: | :---- |
67-
| \-0.1147391 | \-0.0043260 | level\_0 | 0.9387403 | FALSE |
68-
| \-5.8194213 | 0.4713108 | level\_0.5 | \-0.9525914 | FALSE |
69-
| \-8.4565114 | \-0.7274771 | level\_-0.5 | \-0.7777899 | FALSE |
70-
| NA | 0.3300295 | level\_0.5 | 0.0947080 | FALSE |
71-
| NA | 0.1465502 | level\_0 | \-0.1983596 | FALSE |
72-
| NA | 0.6885266 | level\_0.5 | \-0.3045011 | TRUE |
73+
| x | y | xc | x2 | yc |
74+
| ---------: | ----------: | :---------- | ----------: | :---- |
75+
| 1.884861 | 1.0717646 | level\_1 | 0.0046504 | TRUE |
76+
| 1.507742 | 0.9958029 | level\_1 | \-1.2287497 | TRUE |
77+
| \-5.490116 | 0.8315705 | level\_1 | \-0.1405980 | TRUE |
78+
| NA | 0.6007655 | level\_0.5 | \-0.2073270 | TRUE |
79+
| NA | \-0.8339836 | NA | \-0.9215306 | FALSE |
80+
| NA | \-0.5329006 | level\_-0.5 | 0.3604742 | FALSE |
7381

7482
### Some quick data exploration
7583

@@ -80,28 +88,28 @@ unique(d['xc'])
8088
```
8189

8290
## xc
83-
## 1 level_0
84-
## 2 level_0.5
85-
## 3 level_-0.5
86-
## 8 level_1
87-
## 9 <NA>
88-
## 276 level_-1.5
91+
## 1 level_1
92+
## 4 level_0.5
93+
## 5 <NA>
94+
## 6 level_-0.5
95+
## 27 level_0
96+
## 269 level_-1.5
8997

9098
``` r
9199
table(d$xc, useNA = 'always')
92100
```
93101

94102
##
95103
## level_-0.5 level_-1.5 level_0 level_0.5 level_1 <NA>
96-
## 94 1 79 106 113 107
104+
## 94 1 85 98 103 119
97105

98106
Find the mean value of `yc`
99107

100108
``` r
101109
mean(d[['yc']])
102110
```
103111

104-
## [1] 0.318
112+
## [1] 0.324
105113

106114
Plot of `yc` versus `x`.
107115

@@ -182,16 +190,16 @@ knitr::kable(score_frame)
182190

183191
| varName | varMoves | rsq | sig | needsSplit | extraModelDegrees | origName | code | recommended |
184192
| :----------------------------- | :------- | --------: | --------: | :--------- | ----------------: | :------- | :---- | :---------- |
185-
| x | TRUE | 0.0004439 | 0.5983019 | FALSE | 0 | x | clean | FALSE |
186-
| x\_isBAD | TRUE | 0.0000556 | 0.8520676 | FALSE | 0 | x | isBAD | FALSE |
187-
| xc\_catP | TRUE | 0.2293441 | 0.0000000 | TRUE | 5 | xc | catP | TRUE |
188-
| xc\_catB | TRUE | 0.7647009 | 0.0000000 | TRUE | 5 | xc | catB | TRUE |
189-
| x2 | TRUE | 0.0068394 | 0.0386304 | FALSE | 0 | x2 | clean | TRUE |
190-
| xc\_lev\_NA | TRUE | 0.1518186 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
191-
| xc\_lev\_x\_level\_minus\_0\_5 | TRUE | 0.1307155 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
192-
| xc\_lev\_x\_level\_0 | TRUE | 0.1074272 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
193-
| xc\_lev\_x\_level\_0\_5 | TRUE | 0.0128574 | 0.0045745 | FALSE | 0 | xc | lev | TRUE |
194-
| xc\_lev\_x\_level\_1 | TRUE | 0.5486677 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
193+
| x | TRUE | 0.0005756 | 0.5470919 | FALSE | 0 | x | clean | FALSE |
194+
| x\_isBAD | TRUE | 0.0000771 | 0.8255885 | FALSE | 0 | x | isBAD | FALSE |
195+
| xc\_catP | TRUE | 0.0008468 | 0.4652101 | TRUE | 5 | xc | catP | FALSE |
196+
| xc\_catB | TRUE | 0.7883578 | 0.0000000 | TRUE | 5 | xc | catB | TRUE |
197+
| x2 | TRUE | 0.0026075 | 0.2000083 | FALSE | 0 | x2 | clean | FALSE |
198+
| xc\_lev\_NA | TRUE | 0.1750095 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
199+
| xc\_lev\_x\_level\_minus\_0\_5 | TRUE | 0.1328708 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
200+
| xc\_lev\_x\_level\_0 | TRUE | 0.1185254 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
201+
| xc\_lev\_x\_level\_0\_5 | TRUE | 0.0644178 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
202+
| xc\_lev\_x\_level\_1 | TRUE | 0.4701626 | 0.0000000 | FALSE | 0 | xc | lev | TRUE |
195203

196204
Note that the variable `xc` has been converted to multiple variables:
197205

@@ -227,9 +235,7 @@ score_frame[score_frame[['recommended']], 'varName', drop = FALSE] %.>%
227235

228236
| | varName |
229237
| -- | :----------------------------- |
230-
| 3 | xc\_catP |
231238
| 4 | xc\_catB |
232-
| 5 | x2 |
233239
| 6 | xc\_lev\_NA |
234240
| 7 | xc\_lev\_x\_level\_minus\_0\_5 |
235241
| 8 | xc\_lev\_x\_level\_0 |
@@ -242,10 +248,12 @@ score_frame[!score_frame[['recommended']], 'varName', drop = FALSE] %.>%
242248
knitr::kable(.)
243249
```
244250

245-
| varName |
246-
| :------- |
247-
| x |
248-
| x\_isBAD |
251+
| | varName |
252+
| - | :------- |
253+
| 1 | x |
254+
| 2 | x\_isBAD |
255+
| 3 | xc\_catP |
256+
| 5 | x2 |
249257

250258
Notice that `d_prepared` only includes derived variables and the outcome
251259
`y`:
@@ -256,14 +264,14 @@ d_prepared %.>%
256264
knitr::kable(.)
257265
```
258266

259-
| x | x\_isBAD | xc\_catP | xc\_catB | x2 | xc\_lev\_NA | xc\_lev\_x\_level\_minus\_0\_5 | xc\_lev\_x\_level\_0 | xc\_lev\_x\_level\_0\_5 | xc\_lev\_x\_level\_1 | yc |
260-
| ----------: | -------: | --------: | -----------: | ----------: | ----------: | -----------------------------: | -------------------: | ----------------------: | -------------------: | :---- |
261-
| \-0.1147391 | 0 | 0.1586826 | \-12.4285135 | 0.9387403 | 0 | 0 | 1 | 0 | 0 | FALSE |
262-
| \-5.8194213 | 0 | 0.2005988 | 0.4207646 | \-0.9525914 | 0 | 0 | 0 | 1 | 0 | FALSE |
263-
| \-8.4565114 | 0 | 0.1766467 | \-12.5357588 | \-0.7777899 | 0 | 1 | 0 | 0 | 0 | FALSE |
264-
| \-0.4097425 | 1 | 0.2342342 | 0.5695332 | 0.0947080 | 0 | 0 | 0 | 1 | 0 | FALSE |
265-
| \-0.4097425 | 1 | 0.1501502 | \-12.3469806 | \-0.1983596 | 0 | 0 | 1 | 0 | 0 | FALSE |
266-
| \-0.1622745 | 1 | 0.2012012 | 0.4912209 | \-0.3045011 | 0 | 0 | 0 | 1 | 0 | TRUE |
267+
| x | x\_isBAD | xc\_catP | xc\_catB | x2 | xc\_lev\_NA | xc\_lev\_x\_level\_minus\_0\_5 | xc\_lev\_x\_level\_0 | xc\_lev\_x\_level\_0\_5 | xc\_lev\_x\_level\_1 | yc |
268+
| ----------: | -------: | --------: | ----------: | ----------: | ----------: | -----------------------------: | -------------------: | ----------------------: | -------------------: | :---- |
269+
| 1.8848606 | 0 | 0.2102102 | 14.206543 | 0.0046504 | 0 | 0 | 0 | 0 | 1 | TRUE |
270+
| 1.5077419 | 0 | 0.2005988 | 14.139786 | \-1.2287497 | 0 | 0 | 0 | 0 | 1 | TRUE |
271+
| \-5.4901159 | 0 | 0.2005988 | 14.139786 | \-0.1405980 | 0 | 0 | 0 | 0 | 1 | TRUE |
272+
| \-0.1276897 | 1 | 0.1891892 | 1.219475 | \-0.2073270 | 0 | 0 | 0 | 1 | 0 | TRUE |
273+
| \-0.3929879 | 1 | 0.2402402 | \-12.844663 | \-0.9215306 | 1 | 0 | 0 | 0 | 0 | FALSE |
274+
| \-0.2908461 | 1 | 0.1766467 | \-12.563128 | 0.3604742 | 0 | 1 | 0 | 0 | 0 | FALSE |
267275

268276
## A Closer Look at `catB` variables
269277

@@ -520,9 +528,9 @@ d %.>%
520528

521529
| rsq | count | sig | var |
522530
| -----------: | ----: | --------: | :-- |
523-
| 0.0004438531 | 2 | 1.0000000 | x |
524-
| 0.0068394173 | 3 | 0.1158911 | x2 |
525-
| 0.7638059755 | 2 | 0.0000000 | xc |
531+
| 0.0005756197 | 2 | 1.0000000 | x |
532+
| 0.0026074775 | 3 | 0.6000248 | x2 |
533+
| 0.7883476986 | 2 | 0.0000000 | xc |
526534

527535
More on non-linear variable scoring can be found
528536
[here](https://cran.r-project.org/web/packages/vtreat/vignettes/VariableImportance.html).

‎Examples/Classification/Classification_FT_files/figure-gfm/unnamed-chunk-13-1.png‎

104 Bytes
Loading

‎Examples/Classification/Classification_FT_files/figure-gfm/unnamed-chunk-14-1.png‎

-2.54 KB
Loading

‎Examples/Classification/Classification_FT_files/figure-gfm/unnamed-chunk-15-1.png‎

-5.5 KB
Loading

‎Examples/Classification/Classification_FT_files/figure-gfm/unnamed-chunk-16-1.png‎

-6.81 KB
Loading

‎Examples/Classification/Classification_FT_files/figure-gfm/unnamed-chunk-6-1.png‎

2.14 KB
Loading

‎Examples/Classification/Classification_files/figure-gfm/unnamed-chunk-12-1.png‎

62 Bytes
Loading

0 commit comments

Comments
 (0)