Middle and small-sized enterprise dishonesty risk prediction method based on time sequence knowledge graph

Abstract

The invention provides a middle and small-sized enterprise dishonesty risk prediction method based on a time sequence knowledge graph. The method comprises the steps: constructing a time sequence knowledge graph of small and medium-sized enterprises; performing representation learning on the time sequence knowledge graph to obtain a corresponding entity embedding matrix and a relationship embedding matrix; performing vector representation on adjacent nodes of each enterprise in the time sequence knowledge graph according to the entity embedding matrix and the relationship embedding matrix to obtain a corresponding adjacent node sequence, and calculating a hidden state of each adjacent node according to the adjacent node sequence; according to the hidden state and the representation vector of the target enterprise, analyzing the attention weight of each adjacent node, and then performing aggregation processing to obtain a risk assessment representation vector of the target enterprise; finally, inputting the risk assessment representation vector into a full connection layer to calculate and obtain a dishonesty risk assessment result of the target enterprise, and performing visual display. Compared with a traditional machine learning method, the accuracy of middle and small-sized enterprise dishonesty risk prediction is improved.

Description

technical field [0001] The invention relates to the field of Internet technology, in particular to a method for predicting the risk of untrustworthiness of small and medium-sized enterprises based on time series knowledge graphs. Background technique [0002] Generally speaking, the credit risk of an enterprise can be analyzed from various angles. One of the important perspectives is to analyze the risk of credit dishonesty, that is, whether the enterprise is executed by the court as a dishonest entity. This risk has a great impact on many companies, governments, banks and other institutions that have business dealings with dishonest companies. Therefore, it is necessary to analyze risks to help institutions avoid possible losses and invest more resources in companies with better quality. With the development of cognitive computing, a large amount of unstructured data is used to assist decision-making. Key elements of current credit risk prediction models include corporat...

AI Technical Summary

Problems solved by technology

However, it is very difficult to obtain such information for SMEs

Researchers can only collect useful information from news, cases, and corporate branches, and the accuracy of risk assessment is poor

The effect of the existing evaluation method through machine learning is not very good, so it is necessary to provide a solution to improve the accuracy of the prediction of SMEs' dishonesty risk

Images