DNN-anisotropy-model-building
A Deep Neural Network (DNN) trained with synthetics is proposed to predict seismic anisotropy of shales from conventional well logs in our paper submitted to Journal of Geophysical Research (Solid Earth), which will be published soon. We share the pre-trained DNN here. Some necessary data, including the collected laboratory measurements of shale anisotropy and a portion of the field well log data, can be found under the data folder. Try out the commond 'python predict.py' to predict seismic anisotropy in no time!