A sparse Bayesian network incremental learning method based on continuous industrial data

• Main Authors: GAO YANG, GAO JIN, QI XIAOLONG, YANG CHENCHEN, ZHOU CHUNLEI, SUN SHUANZHU, LI DOU, LI CHUNYAN, ZHANG TIANCHENG, SHUAI YUNFENG, PAN MIAO, WANG QIXIANG, ZHANG YOUWEI, WANG MING
• Format: Patent
• Language: Chinese; English
• Published: 11.01.2019
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a sparse Bayesian network incremental learning method based on continuous industrial data. The sparse Bayesian network incremental learning method based on continuous industrial data comprises the following steps: 1, data preprocessing: preprocessing training samples such as regularization to ensure that the mean value is 0 and the variance is 1; 2, structure learning: based on a sliding window, carrying out incremental learning of the structure in the form of batch samples; 3, parameter learning: obtaining the mean value and variance of the mixed Gaussian distributionof each feature under the current state by using the relation coefficient matrix obtained by the structure learning and the regularized sample, namely the continuous Bayesian network parameters; 4, network updating: for the new arrival samples, using the regularization parameters of the population samples for preprocessing, and forming the same size batch samples by sliding windows for updating the structure and parameter