Business document automatic pushing method and system based on intelligent algorithm

Abstract

The invention discloses a business document automatic pushing method and system based on an intelligent algorithm, and the method comprises the steps: S1, collecting business document data, inputting the business document data into a classification model based on a random forest, and obtaining an initial pushing mode of the business document data; s2, extracting semantic features of the business document data and converting the semantic features into vector representation to obtain semantic feature vectors; s3, constructing an improved graph neural network model based on the semantic feature vector, and obtaining a correction factor based on the graph neural network model; and S4, inputting the correction factor and the initial pushing mode into a Bayesian network to obtain an optimized pushing probability, and when the optimized pushing probability is greater than a pushing threshold value, pushing the business document and optimizing the pushing mode of the business document, thereby reducing the influence of local deviation, integrating multi-source data, and improving the pushing efficiency of the business document. In this way, the problem of inaccurate pushing caused by excessive attention to the local feature data can be avoided.

Images