基于可信度的投票列表合并算法

在投票系统中,每个投票人按照自己对候选人的认可程度对候选人进行排名,从而得到大量的有序投票列表.为了从这些列表中得到一个综合投票结果,需要找到一种合理有效的列表合并算法,综合分析列表数据并将它们合并为一个综合列表.本文提出一种基于可信度的投票列表合并算法,其基本思路是:通过综合分析投票列表中蕴含的众多排名信息,度量出每个列表中每条排名信息可被采信的程度,简称为可信度,然后基于已经得到的可信度,让那些高可信度的排名信息在综合排名中发挥更大的作用,从而得到一个更好的综合排名结果.实验结果充分表明,本文提出的算法能够更有效地挖掘出排名信息的可信度,从而得到准确度更高的合并结果.

Credibility-based Algorithm for Merging Vote Lists

In a voting system， each voter makes a preferential list about candidates， thus a large amount of ordered lists are obtained. To get a comprehensive voting result from these lists， an effective lists merging algorithm is required， which can analyze these lists data and output a comprehensive list. A merging algorithm based on credibility is proposed. Through analyzing the data of lists， numerous ranking messages are extracted， then the credibility of them is formulated and measured， with which the final comprehensive list is computed such that those ranking messages with high credibility could play a more influential role in the final ranking result. Experimental results fully indicate that the algorithm proposed can dig out the credibility about ranking information more effectively， thus attaining the merging results more accurately.