Need help with Imbalance-XGBoost?

Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

166 Stars 36 Forks MIT License 53 Commits 6 Opened issues

XGBoost for label-imbalanced data: XGBoost with weighted and focal loss functions

Readme

This software includes the codes of Weighted Loss and Focal Loss [1] implementations for Xgboost 2 in binary classification problems. The principal reason for us to use Weighted and Focal Loss functions is to address the problem of label-imbalanced data. The original Xgboost program provides a convinient way to customize the loss function, but one will be needing to compute the first and second order derivatives to implement them. The major contribution of the software is the drivation of the gradients and the implementations of them.

**The project has been posted on github for several months, and now a correponding API on Pypi is released. Special thanks to @icegrid and @shaojunchao for help correct errors in the previous versions. The codes are now updated to version 0.7 and it now allows users to specify the weighted parameter \alpha and focal parameter \gamma outside the script. Also it supports higher version of XGBoost now.**

**From version 0.7.0 on Imbalance-XGBoost starts to support higher versions of XGBoost and removes supports of versions earlier than 0.4a30(XGBoost>=0.4a30). This contradicts with the previous requirement of XGBoost<=0.4a30. Please choose the version fits your system accordingly.**

Installing with Pypi will be easiest way, you can run:

pip install imbalance-xgboost

If you have multiple versions of Python, make sure you're using Python 3 (run with

pip3 install imbalance-xgboost). Currently, the program only supports Python 3.5 and 3.6.

The package has hard depedency on numpy, sklearn and xgboost.

To use the wrapper, one needs to import *imbalance_xgboost* from module **imxgboost.imbalance_xgb**. An example is given as bellow:

from imxgboost.imbalance_xgb import imbalance_xgboost as imb_xgb

The specific loss function could be set through *special_objective* parameter. Specificly, one could construct a booster with:

Python xgboster = imb_xgb(special_objective='focal')for

Python xgboster = imb_xgb(special_objective='weighted')for

Python from sklearn.model_selection import GridSearchCV xgboster_focal = imb_xgb(special_objective='focal') xgboster_weight = imb_xgb(special_objective='weighted') CV_focal_booster = GridSearchCV(xgboster_focal, {"focal_gamma":[1.0,1.5,2.0,2.5,3.0]}) CV_weight_booster = GridSearchCV(xgboster_weight, {"imbalance_alpha":[1.5,2.0,2.5,3.0,4.0]})The data fed to the booster should be of numpy type and following the convention of:

x: [nData, nDim]

y: [nData,]

In other words, the x

And finally, one could fit the data with Cross-validation and retreive the optimal model:

```Python CV

After getting the optimal booster, one will be able to make predictions. There are following methods to make predictions with imabalnce-xgboost:Python raw

Method `predict`

This will return the value of 'zi' before applying sigmoid.Python sigmoid

Method `predict_sigmoid`

This will return the \hat{y} value, which is p(y=1|x) for 2-lcass classification.predict

Method

Python class_output = opt_focal_booster.predict_determine(data_x, y=None)

This will return the predicted logit, which 0 or 1 in the 2-class scenario.predict

Method

Python prob_output = opt_focal_booster.predict_two_class(data_x, y=None)

This will return the predicted probability of 2 classes, in the form of [nData * 2]. The first column is the probability of classifying the datapoint to 0 and the second column is the prob of classifying as 1.score

To assistant the evluation of classification results, the package provides a score function

with multiple metrics. One can usemake

method in sk-learn andfunctools

to specify the evaluation score. The method will be compatible with sk-learn cross validation and model selection processes.``Python import functools from sklearn.metrics import make

xgboost*opt*param = CV*focal*booster.best*params*

xgboost*opt = imb*xgb(special*objective='focal', **xgboost*opt_param)

loo_splitter = LeaveOneOut()

score*eval*func = functools.partial(xgboost*opt.score*eval_func, mode='accuracy')

loo*info*dict = cross*validate(xgboost*opt, X=x, y=y, cv=loo*splitter, scoring=make*scorer(score*eval*func))

In the new version, we can also collect the information of the confusion matrix through the `correct_eval_func` provided. This enables the users to evluate the metrics like accuracy, precision, and recall for the average/overall test sets in the cross-validation process.Python

TP*eval*func = functools.partial(xgboost*opt.score*eval*func, mode='TP')
TN*eval*func = functools.partial(xgboost*opt.score*eval*func, mode='FP')
FP*eval*func = functools.partial(xgboost*opt.score*eval*func, mode='TN')
FN*eval*func = functools.partial(xgboost*opt.score*eval*func, mode='FN')

score*dict = {'TP': make*scorer(TP*eval*func),
'FP': make*scorer(TN*eval*func),
'TN': make*scorer(FP*eval*func),
'FN': make*scorer(FN*eval_func)}

loo*info*dict = cross*validate(xgboost*opt, X=x, y=y, cv=loo*splitter, scoring=score*dict)
overall*tp = np.sum(loo*info*dict['test*TP']).astype('float')
```
More soring function may be added in later versions.

You don't have to understand the equations if you find they are hard to grasp, you can simply use it with the API. However, for the purpose of understanding, the derivatives of the two loss functions are listed.

For both of the loss functions, since the task is 2-class classification, the activation would be sigmoid:

And bellow the two types of loss will be discussed respectively.

And combining with $\hat{y}$, which are the true labels, the weighted imbalance loss for 2-class data could be denoted as:

Where $\alpha$ is the 'imbalance factor'. And $\alpha$ value greater than 1 means to put extra loss on 'classifying 1 as 0'.

The gradient would be:

And the second order gradient would be:

The focal loss is proposed in [1] and the expression of it would be:

The first order gradient would be:

And the second order gradient would be a little bit complex. To simplify the expression, we firstly denotes the terms in the 1-st order gradient as the following notations:

Using the above notations, the 1-st order drivative will be:

Then the 2-nd order derivative will be:

If you use this package in your research please cite our paper:

@misc{wang2019imbalancexgboost, title={Imbalance-XGBoost: Leveraging Weighted and Focal Losses for Binary Label-Imbalanced Classification with XGBoost}, author={Chen Wang and Chengyuan Deng and Suzhen Wang}, year={2019}, eprint={1908.01672}, archivePrefix={arXiv}, primaryClass={cs.LG} }

@author: Chen Wang, Dept. of Computer Science, School of Art and Science, Rutgers University (previously affiliated with University College London, Sichuan University and Northwestern Polytechnical University)

@version: 0.7.4

[1] Lin, Tsung-Yi, Priyal Goyal, Ross Girshick, Kaiming He, and Piotr Dollár. "Focal loss for dense object detection." IEEE transactions on pattern analysis and machine intelligence (2018).

[2] Chen, Tianqi, and Carlos Guestrin. "Xgboost: A scalable tree boosting system." In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pp. 785-794. ACM, 2016.