A forecasting method based on knowledge map and complex network combination

Abstract

The invention provides a prediction method based on a knowledge map and a complex network combination. The method comprises the following steps: acquiring a plurality of software fault cases of different types; anyalzing Multiple fault phenomena and fault causes in fault cases by clustering analysis. The key words of phenomenon clustering and cause clustering are extracted as clustering labels ofeach category to generate knowledge map. Wherein Clustering labels are respectively corresponding to a plurality of functional modules of the software; Acquiring a mapping relationship between each functional module and the software code; Setting up code network; mapping The functional modules corresponding to the clustering label to the code network under each version, and the corresponding codepart is marked to predict the location of the code network risk of the unknown version software. The invention can effectively mark the specific software fault correspondence into the code network, and then predict the risk of the unknown version of the software code network, and then implement effective risk avoidance measures.

Description

technical field [0001] The invention provides a prediction method based on the combination of knowledge graph and complex network, which belongs to the field of software prediction. Background technique [0002] With the development of science and technology, software is used more and more frequently, which puts forward higher requirements for the accuracy and efficiency of software prediction. Usually, people’s prediction method for software failure is to filter out the measurement elements related to software failure by analyzing software code or development process, and then create a failure prediction data set by mining software history warehouse, and build a failure prediction model to predict The number of potential faults of the software under test and the approximate program location where the fault occurred. [0003] However, the model established by this method can usually only predict the number of software faults and the rough location of the code program corres...

AI Technical Summary

Problems solved by technology

[0003] However, the model established by this method can usually only predict the number of software faults and the rough location of the code program corresponding to the fault. For more detailed fault categories, it is not reflected in the corresponding code program module.

When a failure occurs, relevant technical personnel cannot accurately and adequately improve the code layer, so that they cannot effectively avoid code network risks

Images