基于机器学习的设备故障预测分析方法

数据准备（Data preparation）——数据处理（Merging data sources）——特征工程（Feature engineering: lag feature, static feature）——建模（Modeling: Bin-class, regression, multi-class）——训练、仿真（Training, Simulation）——决策（Decision）

++Binary classification for predictive maintenance: to predict the probability that an equipment will fail within a future time period.

++Regression for predictive maintenance: to compute the remaining

useful life (RUL) of an asset

++Multi-class classification for predictive maintenance: to assign an asset to one of the multiple possible periods of time to give a range of time to failure for each asset, and to identify the likelihood of failure in a future period due to one of the multiple root causes.