a-aron-j / well_analysis

This study aims to investigate the utilization of Artificial Neural Networks (ANN) for the prediction of water distribution in oil reservoirs done in collaboration with Heriot Watts University. The data is taken from log data produced by sensors on drilling pipelines located at the North Sea. The presence of oil is inversely proportional to the presence of water, thus enforcing the importance of this study for the energy sector. In order to further validate the feasibility of applying ANN, the performance of ANN will be compared to another machine learning algorithm, which is Support Vector Regression (SVR). SVR is a modified version of Support Vector Machine (SVM), which is optimized for numerical regression. In contrary, SVM is optimized for categorical classification.