Steps for data prediction
Steps for data prediction:
Data preprocessing (normalization)
(1) outlier handling (for data that obey normal distribution or approximate normal distribution, and the number of measurements is more than 10, Rajda criterion, that is, 3 Sigma criterion, that is, outliers if the deviation is more than 3 Sigma)
(2) normalization (the normal data after eliminating abnormal values are dimensionless by range normalization, and the results are as follows: [0Jing 1]. Data prediction (neural network prediction, or grey prediction, or differential equation prediction)
