Detecting phase transitions is one of the challenging problems in condensed matter physics. For systems, which show phase transitions, in which an order parameter smoothly becomes nonzero, identifying critical points needs finite-size scaling of very large systems. There also exist phase transitions in nature, that the order parameter is not precisely specified. Hence the detection of the phase transitions is a difficult task. Machine Learning methods are supposed to be powerful tools for investigating phase transition. In this thesis, we first introduce the structure of machine learning algorithms and describe the corresponding building blocks. We then introduce neural networks algorithms...
Detecting phase transitions is one of the challenging problems in condensed matter physics. For systems, which show phase transitions, in which an order parameter smoothly becomes nonzero, identifying critical points needs finite-size scaling of very large systems. There also exist phase transitions in nature, that the order parameter is not precisely specified. Hence the detection of the phase transitions is a difficult task. Machine Learning methods are supposed to be powerful tools for investigating phase transition. In this thesis, we first introduce the structure of machine learning algorithms and describe the corresponding building blocks. We then introduce neural networks algorithms...