Novel technique of fusion for truth value based on poor information

Poor information means incomplete and insufficient information for the characteristic presented in the subject investigated. A novel technique of truth value fusion is proposed to solve the problem about parameter estimation in information poor system. This technique contains two approaches: one is that different estimated truth values are obtained via different methods and rules from raw data series and the other is that the final estimated truth value is just the fused value which satisfies the range rule, by using these different methods and rules to repeat the fusion of the fused series from the estimated truth value series. The proposed technique allows the probability distribution unknown and the number of data very small. Experiment on the data series of the measured temperature shows the final estimated truth value is able to be regarded as the rightest deputy of the truth value in the information poor system studied.