Incremental Self-organizing Maps Method of Electric Power Big Data Based on Multi-task Feature Fusion Algorithm
In order to achieve accurate mapping of electric power big data,the incremental self-organizing maps method of elec-tric power big data is studied based on multi-task feature fusion algorithm.The electric power big data are decomposed,and the feature types are divided in the form of linear combinations.The multi-task feature fusion algorithm is used to design the num-ber of candidate classification categories and determine the target of self-organizing maps.The incremental self-organizing maps of electric power big data is achieved by the semantic length of different data corresponding to the types they belong to,and the design of the self-organizing maps method of incremental data is completed.The new method is verified by taking the electric power company which is actually running in a province as the test object,and the electric power big data generated in one year as the test sample,which is mapped according to the specific types respectively.The experimental results show that the new method can achieve accurate self-organizing maps,does not produce data exchange errors in the whole process,and has applica-tion value.
electric power big dataself-organizing mapmulti-task feature fusion algorithmfeature type
万方数据
维普
