A Novel Method Based on Multi-attribute Fusion for Identifying Lead Users in Online Communities
摘要
在线社区已经成为企业与用户进行价值共创的重要平台.社区中的领先用户能够反映市场的潜在需求,为企业的产品创新提供重要信息和改进方向.因此,识别领先用户对企业的产品开发和创新具有重要的价值.为了有效识别出在线社区中的领先用户,本文提出了基于用户行为属性、用户网络位置属性和用户情感属性的多属性融合的在线社区领先用户识别方法.首先,结合领先用户基本理论,明晰在线社区中领先用户的特征.其次,构建了多属性融合的在线社区领先用户量化特征指标.最后,应用朴素贝叶斯(Naive Bayes,NB)算法构建分类器实现对领先用户的识别.结果显示,在考虑多属性融合的在线社区领先用户特征指标的情况下,各分类器对领先用户的识别效果显著提升.相比于其他方法,NB分类器获得了最高的Recall值,平均运行时间最少,验证了本文所提出方法对在线社区领先用户的识别准确性和效率方面的优越性.
Online community has become an important platform for enterprises to collaborate with their customers.The lead users in the online community can reflect the potential demand of the market,providing important information and improvement directions for product innovation of enterprises.Therefore,identifying lead users is of great value to the product development and innovation for enterprises.To effectively identify the lead users in online communities,a multi-attribute fusion method,based on user behavior attributes,user network location attributes and user emotional attributes,is proposed for identifying lead users from online communities.Firstly,based on the fundamental theory of lead users,the characteristics of lead users in online communities are clarified.Secondly,a quantitative indicator system for online community lead users,based on multi-attribute fusion,is constructed.Finally,a classifier is estimated by utilizing Naive Bayes(NB)algorithm to identify lead users.The results indicate that when considering the multi-attribute fusion characteristic indicators of lead users in online communities,the overall performance of each classifier has significantly improved.Compared with other methods,the NB performs returns the highest score of Recall and the least average running time,which verifies the superiority of the proposed method in terms of accuracy and efficiency in identifying leading users in the online community.