Sedimentary Microfacies Identification Using Principal Component Analysis and Neural Network Based on Genetic Algorithm

Due to the high labor cost and measurement difficulties of traditional methods in determination of sedimentary microfacies, principal component analysis (PCA) and back propagation neural network (BPNN) based on genetic algorithm (GA) are employed for the recognition system. PCA can extract the most distinguishing vectors which have been removed noise in the discarded dimensions from the original data and extremely reduce the computational burden. GA can find the optimal weights and biases of the neutral network which will avoid meeting the local optimal value. The experimental results of this identification system show that PCA-GA-BPNN gives superior predictions over ordinary neutral network. More importantly, this method gets rid of the tedious activities, improves efficiency as well as maintains high recognition accuracy and also has significant potential applications in oil exploration and development field.