Method and system for predicting network service quality

Abstract

The invention discloses a network service quality prediction method and system, and the method comprises the steps: obtaining user data and service data, inputting the user data and the service data into a trained network service quality prediction model, and obtaining a network service quality prediction result, wherein the trained network service quality prediction model is obtained by training different sample data and corresponding soft labels, the sample data comprises user training data and service training data, and the soft label is prediction data output by the trained DeepFM model after the sample data is input. According to the method, the DeepFM model is adopted, manual feature combination is not needed, a sparse data set can be processed, the prediction data output by the trained DeepFM model after the sample data is input serves as the soft label to train the network service quality prediction model, the scale of the network service quality prediction model is reduced, and the operation and maintenance load is relieved.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method and system for network service quality prediction. Background technique [0002] In recent years, the explosive development of network services has prompted the rapid expansion of the number of homogeneous network services provided by various network service providers, and it has become a major difficulty to select the most suitable network service for users from a large number of homogeneous network services. Network service quality prediction technology can effectively solve this problem by predicting the quality of a large number of homogeneous network services that have never been used. [0003] In the problem of prediction, it is very important to obtain and learn knowledge information from a large amount of data. Commonly used methods include Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Factorization Machines (FM), Wide&Deep, etc. Among...

AI Technical Summary

Problems solved by technology

Among them, CNN tends to learn the combined features of adjacent features; RNN is suitable for data with a time series relationship; FM uses latent vectors as inner products to represent combined features, which theoretically solves the problem of low-order and high-order combined feature extraction , due to the limitation of computational complexity, it is very limited in practical application; Wide&Deep can learn low-order and high-order combination features at the same time, but relies on artificial feature engineering, the model results are greatly affected by feature engineering, which increases the manpower and calculation Resource requirements, and difficult to deal with sparse data sets and high-dimensional information, not suitable for web service data containing high-dimensional information

Images