The repository is still being updated.
The paper shows the possibility of combining statistical methods in the field of analysing large data sets, the so-called Big Data, in terms of pattern recognition – in this case, the recognition of geological structures. The methods used were: principal component analysis (PCA) and multi-layer perceptron neural networks (MLP). PCA was used to obtain new predictors for MLP networks from primary data (profiling in wells). The network obtained in this way was compared with the network, where the predictors were the original values. It was found that neural networks in combination with PCA, due to the reduction in the number of input data are less complicated in construction, and thus the final calculations run faster with high efficiency. The ability to learn network faster and high precision model predictions give very optimistic results. All statistical analyses were performed in the Python environment.