Improved Uncertainty Distribution of Single Expert Data

Distribution is a basic characteristic of statistics. There exists only two ways to obtain the distribution function for some quantity, which are the frequency generated by historical data and the belief degree evaluated by domain experts, respectively. However, it is undoubtedly dicult for expert to give specific and accurate experimental data in every questionnaire. By improving the questionnaire, this paper proposes a new method of data collection combining uncertainty and randomness. Besides, a method of moments for estimating the distributions with known parameters is estimated by using the collected data. Several numerical examples are provided to illustrate the feasibility of the method.