دانلود رایگان مقاله انگلیسی یک مدل ادغام اطلاعاتی از اتحادیه های نوآوری بر اساس شبکه بیزی - IEEE 2018

Abstract:

To solve the problem of information fusion from multiple sources in innovation alliances, an information fusion model based on the Bayesian network is presented. The multi-source information fusion process of innovation alliances was classified into three layers, namely, the information perception layer, the feature clustering layer, and the decision fusion layer. The agencies in the alliance were defined as sensors through which information is perceived and obtained, and the features were clustered. Finally, various types of information were fused by the innovation alliance based on the fusion algorithm to achieve complete and comprehensive information. The model was applied to a study on economic information prediction, where the accuracy of the fusion results was higher than that from a single source and the errors obtained were also smaller with the MPE less than 3%, which demonstrates the proposed fusion method is more effective and reasonable. This study provides a reasonable basis for decision-making of innovation alliances.

5 Conclusion

An information fusion hierarchy model for innovation alliances based on the Bayesian network was presented. The decision-making process of the innovation alliance is divided into three layers, namely, the information perception layer, the feature clustering layer, and the decision fusion layer. The process of information fusion in the alliance is as follows. Information is observed via sensors in the information perception layer and clustered according to the features in the feature clustering layer. Then, comprehensive and effective information is obtained through the fusion algorithm in the decision fusion layer. The proposed technology presents the advantages of multiple sensors operating jointly, and the accuracy and reliability of the information obtained in the alliance was improved through collaborative work. The fusion algorithm improves the reliability of the fused information for decision-making in the innovation alliance.

The information fusion model was applied to research on predicting economic information, and a reasonable fusion algorithm was presented. Simulation results indicated that the results of fusion were more credible and comprehensive than those of a single-agency. As well, the prediction accuracy of fusion results was higher and errors were smaller.

The information fusion model could process and integrate information from multiple sources, achieve overall optimization of information, and form complete and comprehensive information based on a fusion algorithm. This study provides a theoretical framework of the process of information transfer and integration, as well as scientific support for big data analysis and decision making in innovation alliances.