Multi-source heterogeneous data fusion method and system

Abstract

The invention provides a multi-source heterogeneous data fusion method and system, and relates to the technical field of data management, and the method comprises the following steps: carrying out the recognition and classification of multi-source heterogeneous data, determining the format, structure and content of a data source, and creating a corresponding data source configuration; data preprocessing work is carried out on each data source; data samples are collected and preprocessed to obtain preprocessed data samples; establishing a strategy generation model based on the preprocessed data samples and corresponding data mapping strategies and data type conversion strategies; according to the method, the strategy generation model is established, the fusion strategy can be automatically adjusted according to the characteristics and changes of different data sources, the complexity of multi-source heterogeneous data is adapted, through reinforcement learning, the model can continuously optimize decisions and select the optimal data matching and conversion strategy in a dynamic environment, an enterprise can rapidly integrate data of a new channel, and the efficiency is improved. The competitiveness is improved, and the method adapts to the rapidly changing market environment.

Images