On a Distributed Fusion Algorithm in Oil Forecast

Distributed fusion algorithm and its model (DFM) are discussed for oil forecast in this paper. DFM comprises a global fusion center (GFC) and several local fusion units (LFU) tightly connecting with each other. LFU performs fusion through two steps: the feature-level fusion that analyzes qualitative data through classifying analysis method and extracts quantitative data through BP neural network method; and the decision-level fusion that conducts decision-level analysis on the results of feature-level fusion through Bayesian network. GFC makes the final decision on the LFU results. Experiments proves that DFM is efficient and acceptable since it decreases global complexity by separating one whole fusion tasks into several local fusion ones.