Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

vignette text

Browse Source
This commit is contained in:
pommedeterresautee 2015-02-12 22:44:57 +01:00
parent 276b68b984
commit ae9f7e9307
2 changed files with 17 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
R-package/vignettes/vignette.css
View File

@ -24,7 +24,7 @@ body{
/ color: white;
line-height: 1;
max-width: 960px;
max-width: 800px;
padding: 20px;
font-size: 17px;
}
@ -131,6 +131,8 @@ code {
border-radius: 4px;
padding: 5px;
display: inline-block;
max-width: 800px;
white-space: pre-wrap;
}
code.r, code.cpp {
@ -151,7 +153,7 @@ blockquote {
border-left:.5em solid #606AAA;
background: #F8F8F8;
padding-left: 1em;
margin-left:25px;
margin-left:10px;
max-width: 500px;
}
@ -162,14 +164,14 @@ blockquote cite {
}
blockquote cite:before {
content: '\2014 \00A0';
/content: '\2014 \00A0';
}
blockquote p {
color: #666;
}
hr {
/ width: 540px;
/ width: 540px;
text-align: left;
margin: 0 auto 0 0;
color: #999;

17
R-package/vignettes/xgboostPresentation.Rmd
View File

@ -89,11 +89,15 @@ data(agaricus.train, package='xgboost')
data(agaricus.test, package='xgboost')
train <- agaricus.train
test <- agaricus.test
# Each variable is a S3 object containing both label and data.
```
> In the real world, it would be up to you to make this division between `train` and `test` data. The way you should do it is out of the purpose of this article, however `caret` package may [help](http://topepo.github.io/caret/splitting.html).
Each variable is a `list` containing both label and data.
```{r dataList, message=F, warning=F}
str(train)
```
Let's discover the dimensionality of our datasets.
```{r dataSize, message=F, warning=F}
@ -101,7 +105,7 @@ dim(train$data)
dim(test$data)
```
> Clearly, we have here a small dataset, however **Xgboost** can manage huge one very efficiently.
Clearly, we have here a small dataset, however **Xgboost** can manage huge one very efficiently.
The loaded `data` are stored in `dgCMatrix` which is a *sparse* matrix type and `label` is a `numeric` vector in `{0,1}`.
@ -125,7 +129,7 @@ In *sparse* matrix, cells which contains `0` are not encoded. Therefore, in a da
bstSparse <- xgboost(data = train$data, label = train$label, max.depth = 2, eta = 1, nround = 2, objective = "binary:logistic")
```
> To reach the value of a `S3` object field we use the `$` character.
> To reach the value of a variable in a `list` use the `$` character followed by the name.
Alternatively, you can put your dataset in a *dense* matrix, i.e. a basic *R* matrix.
@ -193,8 +197,9 @@ Here, we have just computed a simple metric: the average error:
* `probabilityVectorPreviouslyComputed != test$label` computes the vector of error between true data and computed probabilities ;
* `mean(vectorOfErrors)` computes the average error itself.
> The most important thing to remember is that to do a classification basically, you just do a regression and then apply a threeshold.
> Multiclass classification works in a very similar way.
The most important thing to remember is that **to do a classification basically, you just do a regression and then apply a threeshold**.
Multiclass classification works in a very similar way.
This metrix is **`r round(err, 2)`** and is pretty low: our yummly mushroom model works well!
@ -282,7 +287,7 @@ watchlist <- list(train=dtrain, test=dtest)
bst <- xgb.train(data=dtrain, max.depth=2, eta=1, nround=2, watchlist=watchlist, objective = "binary:logistic")
```
> **Xgboost** has computed at each round the same average error metric than seen above (we set `nround` to 2, that is why we have two lines of metric here). Obviously, the `train-error` number is related to the training dataset (the one the algorithm learns from) and the `test-error` number to the test dataset.
**Xgboost** has computed at each round the same average error metric than seen above (we set `nround` to 2, that is why we have two lines of metric here). Obviously, the `train-error` number is related to the training dataset (the one the algorithm learns from) and the `test-error` number to the test dataset.
Both training and test error related metrics are very similar, and in some way, it makes sense: what we have learned from the training dataset matches the observations from the test dataset.