The Research And Application Of The Support Vector Machine Algorithm Based On Artificial Intelligence

Data mining is committed to analyzing and understanding the data, as well as revealing the internal relations hidden in the data. It is a process of decision support, mainly based on artificial intelligence, machine learning, pattern recognition, statistics, database and visualization technology. Data mining allows highly automated analysis data, make the inductive reasoning, and dig out the potential patterns, to help policymakers adjust the marketing strategy, reduce the risk and make the right decisions.A number of new concepts, new technologies of data mining appeared after more than ten years of effort of research workers. Especially in recent years, some basic concepts and methods tend to be clearer, and the research of data mining technology is developed towards a higher level. Data mining technology is usually divided into two categories, one is statistical model, the most common used technology of statistical model are probability analysis, relevance, clustering analysis and discriminant analysis, etc. The other is the machine learning in artificial intelligence, getting the needed model and parameter by training and learning a large number of samples set. The choice of data mining technology will influence the quality of the final result, so it usually combines multiple technologies that needed to form a perfect data mining technology.As an important research subject of data mining, forecasting is one of the hot topics of people, and it has been widely used in the social economy and engineering technology. Electric load forecasting plays important role especially in power system. Electric load data is usually collected to build a data warehouse, and people use the data warehouse to forecast the future electric load. But the electric load curve is non-linear, and it is related to many factors which are more and more and change very fast. It is a great important problem that how to process the large amounts of data effectively.In this paper, a new combined forecasting method(ESPLSSVM) based on empirical mode decomposition, seasonal adjustment, particle swarm optimization (PSO) and least squares support vector machine (LSSVM) model is proposed. In the electric market, noise signals usually affect the forecasting accuracy, which are caused by different erratic factors. First of all, ESPLSSVM uses an empirical mode decomposition-based signal filtering method to reduce the influence of noise signals. Secondly, ESPLSSVM eliminates the seasonal components from the de-noised resulting series and then it models the resultant series using the LSSVM which is optimized by PSO (PLSSVM). Here in order to increase the accuracy, instead of using the original LSSVM, the parameters of LSSVM are optimized by PSO.Through the simulation of the electric load in New South Wales of Australia, and make comparison between our proposed method and other three forecasting methods, the following result can be got:our proposed forecasting method effectively improves the electric load forecasting accuracy.