Hybrid Support Vector Machine Model For Intelligent Decision Of High Uncertain Data Sets

Many information systems in the world today produce a large amount of raw data which are often highly uncertain data with characters of highly unbalanced,high-dimensional,large-scale,and various attribute values.In recent years,it has become particularly common to conduct big data mining through reliable machine learning techniques.It is difficult to achieve satisfactory results to mine such high uncertainty data by general machine learning methods.Therefore,in order to excavate more hidden and useful information in high-indefinite data,it is necessary to develop a high-performance machine learning method for processing various highly uncertain data.This paper proposes a hybrid support vector machine model to mine this type of high uncertainty data.Its structure includes preprocessing technology,model establishment and parameter optimization.Data preprocessing includes normalization and data resampling to smooth data based on prior knowledge.By selecting frequency of repeated sampling,it can be realized to make the sample smoother and further improve the classification performance.The established algorithm model is Support Vector Machine(SVM)model.It is a method based on the structural risk minimization principle.It can handle high-dimensional data well and has good generalization ability,but it is only suitable for small sample data.So this paper proposes a method for processing large-scale classification data base on support vector machines to improve the ability of hybrid model handling large-scale data.In this paper,Genetic Algorithm(GA),Particle Swarm Optimization(PSO)and Grid Search(GS)are used to optimize the parameters of the model to improve the performance of the hybrid model.The paper will obtain multiple hybrid support vector machine models by selecting different methods or algorithms during three stages of preprocessing technology,model establishment and parameter optimization,and selects the optimal hybrid support vector machine model through certain strategies in all the mixed model types.This model can handle the high uncertain data.Finally,it was applied to the ECG data set in the medical database and compared with other methods to verify its superiority.