Weighted RF-BiLSTM-based public place personnel health risk prediction method and system

The invention relates to the technical field of public safety, and provides a weighted RF-BiLSTM-based public place personnel health risk prediction method and system. The method comprises the following steps: performing data fusion and feature engineering processing on acquired multi-source heterogeneous data to establish a health risk database suitable for public places with different time scales; based on a weighted random forest method, performing feature selection on the environmental risk indexes in the health risk database to obtain a dimensionality reduction feature set; on the basis of a deep learning framework, according to time series data corresponding to environmental risk indexes in dimension reduction feature sets of different time scales and statistical features contained in a health risk database, public place personnel health risk assessment early warning models of different prediction time scales of a weighted RF-based bidirectional long-short-term memory network are constructed; and predic