基于動(dòng)態(tài)加權(quán)網(wǎng)絡(luò)凝聚度的交互式評(píng)價(jià)信息集結(jié)法

摘要： 針對(duì)多階段交互式群體評(píng)價(jià)信息集結(jié)問(wèn)題,以動(dòng)態(tài)交互網(wǎng)絡(luò)為切入點(diǎn),提出一種新的、基于動(dòng)態(tài)加權(quán)網(wǎng)絡(luò)凝聚度的交互式評(píng)價(jià)信息集結(jié)法。該方法將評(píng)價(jià)者視為節(jié)點(diǎn),以兩兩評(píng)分間的相似性為聯(lián)系,從而構(gòu)造有權(quán)無(wú)向的網(wǎng)絡(luò)鄰接矩陣,并借助網(wǎng)絡(luò)凝聚度來(lái)確定評(píng)價(jià)者權(quán)重,以此"橫向"集結(jié)評(píng)價(jià)信息;通過(guò)緊密度的變化來(lái)確定交互終止判定條件,并設(shè)計(jì)規(guī)劃函數(shù)來(lái)求解階段權(quán)重,進(jìn)而"縱向"合成評(píng)價(jià)結(jié)果。該方法可較全面、準(zhǔn)確地集結(jié)群體意見(jiàn),符合SR原則,同時(shí)評(píng)價(jià)結(jié)果的穩(wěn)定性亦較理想,實(shí)例分析表明了其有效性和可行性。
For the problem of information aggregation for multi-phase interactive group evaluation,this paper takes dynamic interactive network as a starting point,presenting a new method of information aggregation in interactive evaluation based on dynamic weighted network coherence.Evaluators are regarded as nodes,and the similarity between two scores given by two evaluators is the connection,hence a network adjacency matrix of weighted and non-directional is constructed.We determine the weights of evaluators by means of network coherence,on this account we aggregate information"horizontally",giving interactive termination condition combined with the change of closeness.A programming function is devised to calculate the phase weights,"vertically"synthesizing the evaluation results.The method can aggregate group opinions more comprehensively and accurately,which is in line with S-R principle that the stability of evaluation results is also ideal.A case study shows the effectiveness and feasibility of the method.