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

Merge pull request #171 from pommedeterresautee/master

Browse Source 
Vignette (1 updated, 1 new)
This commit is contained in:
Tong He 2015-02-10 16:19:54 -08:00
commit 292f4f0e0d
3 changed files with 281 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
R-package/vignettes/discoverYourData.Rmd
View File

@ -5,7 +5,6 @@ output:
css: vignette.css
number_sections: yes
toc: yes
date: "Wednesday, January 28, 2015"
---
Introduction
@ -13,7 +12,7 @@ Introduction
The purpose of this Vignette is to show you how to use **Xgboost** to discover and better understand your own dataset.
You may know **Xgboost** as a state of the art tool to build some kind of Machine learning models. It has been used to win several [Kaggle](http://www.kaggle.com/) competition ([more information](https://github.com/tqchen/xgboost)).
You may know **Xgboost** as a state of the art tool to build some kind of Machine learning models. It has been [used](https://github.com/tqchen/xgboost) to win several [Kaggle](http://www.kaggle.com) competition.
During these competition, the purpose is to make prediction. This Vignette is not about showing you how to predict anything. The purpose of this document is to explain *how to use **Xgboost** to understand the link between the features of your data and an outcome*.
@ -25,7 +24,7 @@ require(Matrix)
require(data.table)
if (!require(vcd)) install.packages('vcd')
```
*Note that **VCD** is used for one of its embedded dataset only (and not for its own functions).*
> **VCD** is used for one of its embedded dataset only (and not for its own functions).
Preparation of the dataset
==========================
@ -36,8 +35,8 @@ Sometimes the dataset we have to work on have *categorical* data.
A *categorical* variable is one which have a fixed number of different values. By exemple, if for each observation a variable called *Colour* can have only *red*, *blue* or *green* as value, it is a *categorical* variable.
In **R**, *categorical* variable is called `factor`.
Type `?factor` in console for more information.
> In **R**, *categorical* variable is called `factor`.
> Type `?factor` in console for more information.
In this demo we will see how to transform a dense dataframe with *categorical* variables to a sparse matrix before analyzing it in **Xgboost**.
@ -62,18 +61,21 @@ Now we will check the format of each column.
str(df)
```
2 columns have `factor` type, one has `ordinal` type (`ordinal` variable is a categorical variable with values wich can be ordered, here: `None` > `Some` > `Marked`).
> 2 columns have `factor` type, one has `ordinal` type.
> `ordinal` variable is a categorical variable with values wich can be ordered
> Here: `None` > `Some` > `Marked`.
Let's add some new categorical features to see if it helps.
Of course these feature are highly correlated to the Age feature. Usually it's not a good thing in ML, but tree algorithms (including boosted trees) are able to select the best features, even in case of highly correlated features.
For the first feature we create groups of age by rounding the real age. Note that we transform it to `factor` so the algorithm treat them as independant values.
```{r}
df[,AgeDiscret:= as.factor(round(Age/10,0))][1:10]
```
> For the first feature we create groups of age by rounding the real age.
> Note that we transform it to `factor` so the algorithm treat them as independant values.
Following is an even stronger simplification of the real age with an arbitrary split at 30 years old. I choose this value **based on nothing**. We will see later if simplifying the information based on arbitrary values is a good strategy (I am sure you already have an idea of how well it will work!).
```{r}
@ -99,7 +101,7 @@ The purpose is to transform each value of each *categorical* feature in a binary
For example, the column Treatment will be replaced by two columns, Placebo, and Treated. Each of them will be *binary*. For example an observation which had the value Placebo in column Treatment before the transformation will have, after the transformation, the value 1 in the new column Placebo and the value 0 in the new column Treated.
Formulae `Improved~.-1` used below means transform all *categorical* features but column Improved to binary values.
> Formulae `Improved~.-1` used below means transform all *categorical* features but column Improved to binary values.
Column Improved is excluded because it will be our output column, the one we want to predict.
@ -133,8 +135,9 @@ You can see plenty of `train-error: 0.XXXXX` lines followed by a number. It decr
A model which fits too well may [overfit](http://en.wikipedia.org/wiki/Overfitting) (meaning it copy paste too much the past, and is not that good to predict the future).
Here you can see the numbers decrease until line 7 and then increase. It probably means I am overfitting. To fix that I may reduce the number of rounds to `nround = 4`. I will let things like that because I don't really care for the purpose of this example :-)
> Here you can see the numbers decrease until line 7 and then increase.
> It probably means I am overfitting. To fix that I may reduce the number of rounds to `nround = 4`.
> I will let things like that because I don't really care for the purpose of this example :-)
Feature importance
==================
@ -149,7 +152,8 @@ importance <- xgb.importance(sparse_matrix@Dimnames[[2]], model = bst)
print(importance)
```
The column `Gain` provide the information we are looking for.
> The column `Gain` provide the information we are looking for.
> As you can see, features are classified by `Gain`.
`Gain` is the improvement in accuracy brought by a feature to the branches it is on. The idea is that before adding a new split on a feature X to the branch there was some wrongly classified elements, after adding the split on this feature, there are two new branches, and each of these branch is more accurate (one branch saying if your observation is on this branch then it should be classified as 1, and the other branch saying the exact opposite, both new branch being more accurate than the one before the insertion of the feature).
@ -157,8 +161,6 @@ The column `Gain` provide the information we are looking for.
`Frequence` is a simpler way to measure the `Gain`. It just counts the number of times a feature is used in all generated trees. You should not use it (unless you know why you want to use it).
As you can see, features are classified by `Gain`.
Plotting the feature importance
-------------------------------
@ -170,6 +172,9 @@ xgb.plot.importance(importance_matrix = importance)
Feature have been automatically divided in 2 clusters: the interesting features... and the others.
> Depending of the case you may have more than two clusters.
> Default value is to limit them to 10, but you can increase this limit. Look at the function documentation for more information.
According to the plot above, the most important feature in this dataset to predict if the treatment will work is :
* the Age;
@ -177,8 +182,6 @@ According to the plot above, the most important feature in this dataset to predi
* the sex is third but already included in the not interesting feature ;
* then we see our generated features (AgeDiscret). We can see that their contribution is very low.
*Note: Depending of the case you may have more than two clusters. Default value is to limit them to 10, but you can increase this limit. Look at the function documentation for more information.*
Does these results make sense?
------------------------------
@ -224,4 +227,17 @@ Linear model may not be that strong in these scenario.
#xgb.plot.tree(sparse_matrix@Dimnames[[2]], model = bst, n_first_tree = 1, width = 1200, height = 800)
```
Special Note: What about Random forest?
=======================================
As you may know, [Random Forest](http://en.wikipedia.org/wiki/Random_forest) algorithm is cousin with boosting and both are part of the [ensemble leanrning](http://en.wikipedia.org/wiki/Ensemble_learning) family.
Both trains several decision trees for one dataset. The *main* difference is that in Random Forest, trees are independant and in boosting tree N+1 focus its learning on what has no been well modeled by tree N (and so on...).
This difference have an impact on a corner case in feature importance analysis: the *correlated features*.
Imagine two features perfectly correlated, feature `A` and feature `B`. For one specific tree, if the algorithm needs one of them, it will choose randomly (true in both boosting and random forest).
However, in Random Forest this choice will be done plenty of times, because trees are independant. So the **importance** of a specific feature is diluted among features `A` and `B`. So you won't easily know they are important to predict what you want to predict.
In boosting, when as aspect of your dataset have been learned by the algorithm, there is no more need to refocus on it. Therefore, all the importace will be on `A` or `B`. You will know that one of them is important, it is up to you to search for correlated features.

14
R-package/vignettes/vignette.css
View File

@ -130,13 +130,16 @@ aside {
width: 390px;
}
blockquote {
border-left:.5em solid #eee;
padding: 0 1em;
margin-left:0;
max-width: 476px;
font-size:14px;
border-left:.5em solid #606AAA;
background: #f5f5f5;
color:#bfbfbf;
padding: 5px;
margin-left:25px;
max-width: 500px;
}
blockquote cite {
/ font-size:14px;
font-size:14px;
line-height:20px;
color:#bfbfbf;
}
@ -146,7 +149,6 @@ blockquote cite:before {
blockquote p {
color: #666;
max-width: 460px;
}
hr {
/ width: 540px;

241
R-package/vignettes/xgboostPresentation.Rmd Normal file
View File

@ -0,0 +1,241 @@
---
title: "Xgboost presentation"
output:
html_document:
css: vignette.css
number_sections: yes
toc: yes
---
Introduction
============
The purpose of this Vignette is to show you how to use **Xgboost** to make prediction from a model based on your own dataset.
You may know **Xgboost** as a state of the art tool to build some kind of Machine learning models. It has been [used](https://github.com/tqchen/xgboost) to win several [Kaggle](http://www.kaggle.com) competition.
For the purpose of this tutorial we will first load the required packages.
```{r libLoading, results='hold', message=F, warning=F}
require(xgboost)
require(methods)
```
In this example, we are aiming to predict whether a mushroom can be eated.
Learning
========
Dataset loading
---------------
We load the `agaricus` datasets and link it to variables.
The dataset is already separated in `train` and `test` data.
As their names imply, the train part will be used to build the model and the test part to check how well our model works. Without separation we would test the model on data the algorithm have already seen, as you may imagine, it's not the best methodology to check the performance of a prediction (would it even be a prediction?).
```{r datasetLoading, results='hold', message=F, warning=F}
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 loaded data is stored in `dgCMatrix` which is a **sparse matrix** type.
Label is a `numeric` vector in `{0,1}`.
```{r dataClass, message=F, warning=F}
class(train$data)[1]
class(train$label)
```
Basic Training using XGBoost
----------------------------
The most critical part of the process is the training.
We are using the train data. Both `data` and `label` are in each data (explained above). To access to the field of a `S3` object we use the `$` character in **R**.
> label is the outcome of our dataset. It is the classification we want to predict. For these data we already have it, but when our model is built, that is this column we want to guess.
In sparse matrix, cells which contains `0` are not encoded. Therefore, in a dataset where there are plenty of `0`, dataset size is optimized. It is very usual to have such dataset. **Xgboost** can manage both dense and sparse matrix.
```{r trainingSparse, message=F, warning=F}
bstSparse <- xgboost(data = train$data, label = train$label, max.depth = 2, eta = 1, nround = 2, objective = "binary:logistic")
```
Alternatively, you can put your dataset in a dense matrix, i.e. a basic R-matrix.
```{r trainingDense, message=F, warning=F}
bstDense <- xgboost(data = as.matrix(train$data), label = train$label, max.depth = 2, eta = 1, nround = 2,
objective = "binary:logistic")
```
Above, data and label are not stored together.
**Xgboost** offer a way to group them in a `xgb.DMatrix`. You can even add other meta data. It will be usefull for the most advanced features.
```{r trainingDmatrix, message=F, warning=F}
dtrain <- xgb.DMatrix(data = train$data, label = train$label)
bstDMatrix <- xgboost(data = dtrain, max.depth = 2, eta = 1, nround = 2, objective = "binary:logistic")
```
Below is a demonstration of the effect of verbose parameter.
```{r trainingVerbose, message=T, warning=F}
# verbose 0, no message
bst <- xgboost(data = dtrain, max.depth = 2, eta = 1, nround = 2,
objective = "binary:logistic", verbose = 0)
# verbose 1, print evaluation metric
bst <- xgboost(data = dtrain, max.depth = 2, eta = 1, nround = 2,
objective = "binary:logistic", verbose = 1)
# verbose 2, also print information about tree
bst <- xgboost(data = dtrain, max.depth = 2, eta = 1, nround = 2,
objective = "binary:logistic", verbose = 2)
```
Basic prediction using Xgboost
------------------------------
The main use of **Xgboost** is to predict data. For that purpose we will use the test dataset. We remind you that the algorithm has never seen these data.
```{r predicting, message=F, warning=F}
pred <- predict(bst, test$data)
err <- mean(as.numeric(pred > 0.5) != test$label)
print(paste("test-error=", err))
```
> You can put data in Matrix, sparseMatrix, or xgb.DMatrix
Save and load models
--------------------
When your dataset is big, it may takes time to build a model. Or may be you are not a big fan of loosing time in redoing the same thing again and again. In these cases, you will want to save your model and load it when required.
Hopefully for you, **Xgboost** implement such functions.
```{r saveLoadModel, message=F, warning=F}
# save model to binary local file
xgb.save(bst, "xgboost.model")
# load binary model to R
bst2 <- xgb.load("xgboost.model")
pred2 <- predict(bst2, test$data)
# pred2 should be identical to pred
print(paste("sum(abs(pred2-pred))=", sum(abs(pred2-pred))))
```
In some very specific cases, like when you want to pilot **Xgboost** from `caret`, you will want to save the model as a **R** binary vector. See below how to do it.
```{r saveLoadRBinVectorModel, message=F, warning=F}
# save model to R's raw vector
raw = xgb.save.raw(bst)
# load binary model to R
bst3 <- xgb.load(raw)
pred3 <- predict(bst3, test$data)
# pred2 should be identical to pred
print(paste("sum(abs(pred3-pred))=", sum(abs(pred2-pred))))
```
Advanced features
=================
Most of the features below have been created to help you to improve your model by offering a better understanding of its content.
Dataset preparation
-------------------
For the following advanced features, we need to put data in `xgb.DMatrix` as explained above.
```{r DMatrix, message=F, warning=F}
dtrain <- xgb.DMatrix(data = train$data, label=train$label)
dtest <- xgb.DMatrix(data = test$data, label=test$label)
```
Using xgb.train
---------------
`xgb.train` is a powerfull way to follow progress in learning of one or more dataset.
One way to measure progress in learning of a model is to provide to the **Xgboost** a second dataset already classified. Therefore it can learn on the real dataset and test its model on the second one. Some metrics are measured after each round during the learning.
For that purpose, you will use `watchlist` parameter. It is a list of `xgb.DMatrix`, each of them tagged with a name.
```{r watchlist, message=F, warning=F}
watchlist <- list(train=dtrain, test=dtest)
bst <- xgb.train(data=dtrain, max.depth=2, eta=1, nround=2, watchlist=watchlist,
objective = "binary:logistic")
```
> To train with watchlist, we use `xgb.train`, which contains more advanced features than `xgboost` function.
For a better understanding, you may want to have some specific metric or even use multiple evaluation metrics.
`eval.metric` allows us to monitor the evaluation of several metrics at a time. Hereafter we will watch two new metrics, logloss and error.
```{r watchlist2, message=F, warning=F}
bst <- xgb.train(data=dtrain, max.depth=2, eta=1, nround=2, watchlist=watchlist,
eval.metric = "error", eval.metric = "logloss",
objective = "binary:logistic")
```
Manipulating xgb.DMatrix
------------------------
### Save / Load
Like saving models, `xgb.DMatrix` object (which groups both dataset and outcome) can also be saved using `xgb.DMatrix.save` function.
```{r DMatrixSave, message=F, warning=F}
xgb.DMatrix.save(dtrain, "dtrain.buffer")
# to load it in, simply call xgb.DMatrix
dtrain2 <- xgb.DMatrix("dtrain.buffer")
bst <- xgb.train(data=dtrain2, max.depth=2, eta=1, nround=2, watchlist=watchlist,
objective = "binary:logistic")
```
### Information extraction
Information can be extracted from `xgb.DMatrix` using `getinfo` function. Hereafter we will extract `label` data.
```{r getinfo, message=F, warning=F}
label = getinfo(dtest, "label")
pred <- predict(bst, dtest)
err <- as.numeric(sum(as.integer(pred > 0.5) != label))/length(label)
print(paste("test-error=", err))
```
View the trees from a model
---------------------------
You can dump the tree you learned using `xgb.dump` into a text file.
```{r dump, message=T, warning=F}
xgb.dump(bst, with.stats = T)
```
Feature importance
------------------
Finally, you can check which features are the most important.
```{r featureImportance, message=T, warning=F}
importance_matrix <- xgb.importance(feature_names = train$data@Dimnames[[2]], model = bst)
print(importance_matrix)
xgb.plot.importance(importance_matrix)
```