关键词: 综合评价, 群体评价, 情感过滤, 行为数据, 随机模拟

Abstract:

The evaluator's emotional behavior belongs to recessive behavior， such as the degree of worship and love of the evaluated object， which is difficult to quantify. Even the emotional behavior data obtained through a questionnaire survey may be difficult to distinguish between true and false. Aiming at the problem that it is difficult to quantify intangible emotional behaviors in the field of non-self-participatory group evaluation， an objective filtering method for group evaluators’ emotional behavioral data is proposed using the idea of probability.Firstly， in the case of insufficient or unavailable data in the early stage， it is almost impossible to realize and convince the absolute recognition of the existence or nonexistence of intangible emotions. Therefore， the emotional behavior probabilities of the evaluators are calculated based on the absolute， relative， and variation dimensions， so as to filter the data. Secondly， there are relatively many subjective factors involved in emotional behavior data filtering， which will make it challenging to obtain the evaluator's true value in the method verification process. Thus， the random simulation algorithm is used to generate the virtual real value of the evaluators， and the emotional behavior of the group evaluators is inversely verified and analyzed. The trend of the verification parameters in different situations is counted to demonstrate the reliability and effectiveness of this method. Finally， an example of performance appraisal is given to solve the above research content and explain its feasibility.The main innovation and application value of this study are reflected in the following aspects　（1） The original data that cannot determine the existence or nonexistence of emotional relationships is directly converted into the form of emotional probability. And the comprehensive probabilities of the evaluator's possible emotional behavior from multiple dimensions are analyzed to achieve emotional filtering of the original data. （2） Considering that the real value of the evaluation is difficult to obtain， the random simulation method is used to generate the virtual real value of the evaluators to improve the effectiveness and feasibility of the reverse simulation analysis. （3） This method is an objective optimization method， which is easy to operate and does not need other subjective data intervention. It can be used to quickly solve such problems， guide the evaluation makers to design a more scientific evaluation process， and obtain more accurate evaluation conclusions.

Key words: comprehensive evaluation, group evaluation, affective filter, behavioral data, stochastic simulation