Large-Scale Support Vector Machine Classification Algorithm Based on Granulation Mechanism
چکیده
Support vector machine classification algorithm has been deeply studied in the field of intelligent information processing with its solid theoretical foundation and excellent performance, it widely used real life, such as face recognition, text classification, sentiment analysis, spam filtering, etc. The time complexity classic support is proportional to square data size, this poses a serious challenge calculability, effectiveness efficiency for massive data. In paper, large-scale based on granulation mechanism proposed from fusion information. A firstly constructed using binary tree label information, well core sample set obtained by combining instance selection method, then remaining samples construct multiple weak classifiers; finally draw idea granularity generate strong classifier. experimental results standard verify algorithm, provides new research ideas methods