aurorlhc/property_test

Predicting Young's Modulus of porous materials using CNN

Physical properties prediction of heterogeneous materials using supervised learning

Physical Properties: Young's Modulus, Diffusion Coefficient, Permeability Material Samples: Sandstone porous materials with different volume fractions Supervised Learning Algorithm: Convoluted Nerual Network (CNN): Nine layers with six convolution layer and three fully connected layer Residual Network (ResNew): Regression: Linear Regression(RidgeCV), Support Vector Machine(SVR), Multi-layer Perceptron(MLP), K-Nearest Neighbours(KNN), Decision Tree, Random Forest, AdaBoost, Gradient Boost, Bagging