Well Log Data Inversion using Higher Order Neural Networks

We use the multilayer perceptron for well log data inversion. The gradient descent method is used in the back propagation learning rule. The input of the network is the apparent conductivity (Ca) and the output of the network is the true formation conductivity (Ct). The original and the higher order features are used for the training process. According to our experimental results, the expanding higher order input features can get a fast training and a smaller error between the desired output and the actual output. The network with 10 input nodes and expanding the input features to third order, 8 hidden nodes, 10 output nodes, can get the smallest average mean absolute error on simulated well log data. Then, we apply the network to the real field data.