A method for dynamic real-time mining of software process activities from svn log event streams

Abstract

The invention belongs to the field of error detection; error correction; monitoring technology, and discloses a method for dynamic and real-time mining of software process activities from SVN log event streams, and preprocessing of software process logs: extracting each event from a software process log file The path and behavior of the two attribute contents; the attribute content expressed in the form of natural language is structured; the structured data obtained is used as a corpus for training, and the vectorization of the data based on semantic features is completed; and then based on the vectorization The data is clustered according to the distance of the data, and the optimal number of clusters is determined by differential quadratic derivation; finally, the obtained clustering results are used as training samples to construct a classifier to complete the mapping of events to activity categories. The solution of the invention makes the initial label division relatively reasonable; the method of constructing the classifier by using the naive Bayesian algorithm constructs the classifier to process brand new data more stably, and completes the effective mapping of the active label of the software process log.

Description

Technical field [0001] The present invention pertains to error detection; error correction; monitoring technology, and more particularly relates to a dynamic real-time activity process mining software from SVN log event flow method. Background technique [0002] Currently, there is such a commonly used in the industry. Quality of software products depends largely on the use of the product development process. The mid-1980s to the 1990s, carried out in the field of software engineering and software process improvement. Traditional software process, divided into two categories: The first category is the software process assessment and improvement models. The second category is Software Process Modeling. Among them, the software can guide the software development process modeling. Because the software is the reaction of the projection objective things in the computer world, and objective things are evolving, software process must also be in constant dynamic change. Corresponding va...

AI Technical Summary

Problems solved by technology

[0003] In summary, the problems in the prior art are: The uniqueness of the software process log in the traditional process mining is that it has only one case, and the events contained in the case lack activity attributes; the technical problems caused by the lack of activity information: the lack of activity attributes will lead to the failure of process mining, the existing process The mining method is to generate a partial order sequence of activities through the activity information of the process log, that is, the trajectory, further divide the sequence formed by the activity information, and then generate a structured process model such as a petri net

Images