Privacy-protected distributed linear discriminant analysis method, system and electronic device

Abstract

The application provides a privacy-protected distributed linear discriminant analysis method and system and electronic equipment, and relates to the technical field of data processing. The method comprises the following steps: determining global mean data of multiple demand ends according to mean data of local data of the multiple demand ends through a management end; wherein the mean data comprises a sample mean vector and a sample quantity of each category label; determining a first type scatter matrix based on the global mean data and the local data through a demand end; determining a second type scatter matrix based on the global mean data and the received multiple first type scatter matrices through the management end; determining a target projection matrix based on the second type scatter matrix and a preset target dimension of the multiple demand ends through the management end; performing dimension reduction processing on the local data according to the received target projection matrix through the demand end to obtain target data; and processing the target data according to an analysis demand through the demand end to obtain an analysis result.