Service modeling method fusing word embedding and non-negative matrix factorization technologies in cloud computing mode

Abstract

A service modeling method fusing word embedding and non-negative matrix factorization technologies in a cloud computing mode comprises the following steps of 1, counting word frequency information, namely the number of occurrence times of words, in each Mashup service and creating a document-word frequency relation matrix D; 2, counting word co-occurrence information so as to calculate SPPMI matrix information; and 3, based on the first step and the second step, obtaining a word frequency information matrix D of the words of the Mashup service document and a context SPPMI matrix M of the words, decomposing the M to obtain a word embedding information matrix, further combining the two kinds of information, and calculating theme information of the service. Organic unification of the model and non-negative matrix factorization can be realized, and the problem of sparse feature information of the Mashup service can be relieved by introducing word embedding information, so that the Mashup service is effectively modeled.

Description

technical field [0001] The invention relates to a service modeling method integrating word embedding and non-negative matrix decomposition technology in a cloud computing mode Background technique [0002] Driven by the development of cloud computing and the idea of ​​"service-oriented" service computing, more and more companies publish data, resources or related businesses to the Internet in the form of Web services, in order to improve the utilization rate of information and their own competitiveness. . However, traditional SOAP-based Web services have problems such as complex technical systems and poor scalability, making it difficult to adapt to complex and changeable application scenarios in real life. In order to overcome the problems caused by traditional services, in recent years, a lightweight information service combination mode has emerged on the Internet - Mashup technology, which can mix and match a variety of different Web APIs and develop a variety of new Web...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the existing Mashup service description text is short and difficult to model due to sparse features, the present invention proposes a service modeling method that combines word embedding and non-negative matrix decomposition technology, and solves word embedding information based on decomposing the SPPMI matrix. Compared with the neural network model, this method can be organically unified with Non-negative Matrix Factorization (NMF) in the model, and by introducing word embedding information, it can alleviate the problem of sparse feature information of Mashup services, thus effectively analyzing Mashup service for modeling

Images