The problem of diagnosing and forecasting reservoir sensitivity damage timely and accurately is always an important field of the reservoir protection research, this paper based on collecting data from core analysis .the main factors affecting reservoir sensitivity are obtained with single variable regression method, established a model of reservoir sensitivity prediction by applying an approach based on multi-level transfer function quantum neural networks. It effectively improved networks convergence and prediction accuracy. The analysis indicates that the model needs fewer parameters, has wider applicability and reliable results (the coincidence rate attains over 91%).and quantitatively reflect reservoir potential sensitivity, thus it can provide reliable basis to reservoir protection.
Multi-Level Feature Abstraction from Convolutional Neural Networks for Multimodal Biometric Identification