Data Predicting In CNET Using Time-Series-Theory Based GA-BP Network

Data predicting is becoming more and more useful and important. In communication network fields, the colleting, maintaining, analyzing, and especially, predicting on data, are important dally works. In this paper, the data to predicting is collected from the capability analyzing system of a Mobile enterprise, and is used to reflect the business capabilities in multidimensional points of view. The analysis and prediction for this kind of data could give out good advises on the management and decision for the enterprisie.Time Series Analysis theory was proposed early in the 20th century. Researches, such as the nature, the features of time series were deeply and widely discussed. An all-around theory was proposed; and a few methods for data analyzing, such as ARMA, were created.During post decades, the theory of machine learning was developing again and again. And since the last years of 20th century, Artificial Neural Network and SVM had been using for data predicting too. However, the research for data predicting using machine learning are almost focus on the learning algorithms them selves. Compared to Time Series Analyzing, the methods of Machine Learning could analog any nonlinear phenomena and get a precisely result.