Abstract: Facing with specific data reduction method, the users have different evaluations on effectiveness of reduction. In view of the current data reduction effectiveness evaluation method is lack of pertinence and ignoring the personalized needs of users, using the existing index system, an evaluation method of data reduction effectiveness based on user’s interest degree is proposed. The method defines the weight baryon space of the user interest degree, calculates the user’s preference for the index, according to the historical data of the user’s evaluation of data set reduction. The acceptable degree for effectiveness of reduction is calculated approximately using the Monte Carlo simulation, as a parameter to evaluate the reduction effect. It implements the method about effectiveness of data reduction for users, and provides a quantitative basis for recommending data reduction method for system users focusing on different indicators.

Key words: data reduction, effectiveness evaluation, user interest degree, Monte Carlo simulation

摘要： 面对特定数据归约方案，不同的用户对归约效果具有不同的评价，针对目前数据归约效果评估方法缺乏针对性、忽视用户个性化需求的问题，基于现有的指标体系，提出基于用户兴趣度的数据归约效果评估方法。通过定义用户兴趣度的权重子空间，根据用户对数据集归约方案评价的历史数据计算用户对指标的偏好，利用蒙特卡洛仿真近似计算出用户对数据归约效果的可接受程度作为评估归约效果的参数，实现面向不同用户的数据归约效果评估方法，为面向不同关注点的系统用户推荐不同数据归约方案提供定量依据。

关键词: 数据归约, 效果评估, 用户兴趣度, 蒙特卡洛仿真