AdTracker

AdTracking Fraud Detection - detect fraudulent click traffic for ads using

• Deep Learning (ResNet with 1D Convolution)
• Gradient Boosting (LightGBM)

Kaggle: https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection

Usage

from model import LightGBM, ResNet
import os

datapath = '..{}data'.format(os.sep)
trainfile = '{}{}sample_train_data.csv'.format(datapath, os.sep)
valfile = '{}{}sample_val_data.csv'.format(datapath, os.sep)

# example using ResNet model
resnet = ResNet()
resnet_model = resnet.train(trainfile, valfile)
# use trained model for prediction - resnet_model.predict()

# example using LightGBM model
lightgbm = LightGBM()
lightgbm_model = lightgbm.train(trainfile)
# use trained model for prediction - lightgbm_model.predict()