Skip to main content

XBNet is an open source project which is built with PyTorch that works as a Boosted neural network for tabular data

Project description

XBNet - Xtremely Boosted Network

Boosted neural network for tabular data

PWC PWC PWC PWC PWC

XBNET that is built on PyTorch combines tree-based models with neural networks to create a robust architecture that is trained by using a novel optimization technique, Boosted Gradient Descent for Tabular Data which increases its interpretability and performance. Boosted Gradient Descent is initialized with the feature importance of a gradient boosted tree, and it updates the weights of each layer in the neural network in two steps:

  • Update weights by gradient descent.
  • Update weights by using feature importance of a gradient boosted tree in every intermediate layer.

Features

  • Better performance, training stability and interpretability for tabular data.
  • Easy to implement with rapid prototyping capabilities
  • Minimum Code requirements for creating any neural network with or without boosting

Comparison with XGBOOST

XBNET VS XGBOOST testing accuracy on different datasets with no hyperparameter tuning

Dataset XBNET XGBOOST
Iris 100 97.7
Breast Cancer 96.49 96.47
Diabetes 78.78 77.48
Titanic 79.85 80.5
German Credit 71.33 77.66

Installation :

pip install --upgrade git+https://github.com/tusharsarkar3/XBNet.git

Example for using

import torch
import numpy as np
import pandas as pd
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split
from XBNet.training_utils import training,predict
from XBNet.models import XBNETClassifier
from XBNet.run import run_XBNET

data = pd.read_csv('test\Iris (1).csv')
print(data.shape)
x_data = data[data.columns[:-1]]
print(x_data.shape)
y_data = data[data.columns[-1]]
le = LabelEncoder()
y_data = np.array(le.fit_transform(y_data))
print(le.classes_)

X_train,X_test,y_train,y_test = train_test_split(x_data.to_numpy(),y_data,test_size = 0.3,random_state = 0)
model = XBNETClassifier(X_train,y_train,2)

criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)

m,acc, lo, val_ac, val_lo = run_XBNET(X_train,X_test,y_train,y_test,model,criterion,optimizer,32,300)
print(predict(m,x_data.to_numpy()[0,:]))

Output images

Reference

If you make use of this software for your work, we would appreciate it if you would cite us:

@misc{sarkar2021xbnet,
      title={XBNet : An Extremely Boosted Neural Network}, 
      author={Tushar Sarkar},
      year={2021},
      eprint={2106.05239},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Features to be added :

  • Metrics for different requirements
  • Addition of some other types of layers

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Built Distribution

XBNet-1.4.6-py3-none-any.whl (13.3 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page