决策支持系统计算偏差的修正策略研究——以招聘支持系统为例

Research on the Correction Strategy of Calculation Deviation in Decision Support Systems:Taking Recruitment Support System as Example

[研究目的]“计算偏差”是利用决策资源通过计算机做出的决策结果与实际结果之间的差异,尤其在决策支持系统中会产生更严重的负面影响。[研究方法]从招聘支持系统计算偏差入手,分析其表现形式与产生原因,从而构建对招聘支持系统计算偏差的修正策略,并对其进行检验说明,设计效果评价指标。[研究结论]招聘支持系统乃至应用“画像”的决策支持系统需要始终注意防范数据不准、数据与算法模型不匹配两方面问题引起的“计算偏差”,灵活组合识别与过滤虚假信息、压缩情感导向明显的信息、增设算法公平性衡量指标等修正方案以构建修正策略,有利于情报实践在提供便利的同时,抵消其由于自动计算带来的损害,并为其他决策支持系统修正所遇到的“计算偏差”问题提供启示与参考。

[Research purpose]"Calculation deviation"refers to the difference between the decision results made by computer using decision resources and the actual results,and it may have serious negative impacts especially in the decision support system.[Research method]This paper starts from the calculation deviation of recruitment support systems,and analyzes its manifestation and causes,so as to build a correction strategy for the calculation deviation of the recruitment support system.[Research conclusion]Decision support systems for recruitment,as well as other decision support systems applying user profiles should be protected from calculation deviation caused by both data inaccuracy problem and data-model mismatch problem.Smartly combining identifying and filtering disinformation,compressing emotional information,increasing algorithmic fairness evaluation index,and so on,for corrective strategy,can draw on the advantages and avoid disadvantages of information practice,and provide implications for correcting calculation deviation in other decision support systems.