基于大数据的用户画像构建方法与运用

目的以大数据环境为研究基础,为产品研发初期获取准确性较高的用户画像,优化其构建方法。方法基于定性与定量综合集成方法的三层次理论框架,首先通过定量研究挖掘线上社交平台的多模态用户数据并对其进行文本分析,提取具有高影响的用户属性因子。其次,使用优化模糊C—均值聚类算法(FCM算法)对用户属性因子进行聚类分析,获得几类用户属性原型。将其原型作为构建画像的维度框架,通过线下定性研究方法对样本用户进行调研,挖掘深层次需求动机。最终,构建出几类用户画像原型。结论以中国手机年轻用户相机拍照行为研究对象,通过线上社交平台和电商平台的公开数据获取用户拍照相关的评论信息,对其进行文本挖掘和分析,并对分析后的数据进行提取、聚类和定性研究,从而获得3类不同拍照类型的用户画像模型。

Construction Method and Application of Personas Based on Big Data

The work aims to acquire accurate user portraits at the initial stage of product development with the big data environment as the research basis, and optimize its construction methods. Based on the three-level theoretical framework of the qualitative and quantitative comprehensive integration method, the multi-modal user data of online social platforms were firstly mined through quantitative research and text analysis to extract the user attributes with high impact. Secondly, fuzzy c-mean clustering algorithm (FCM algorithm) was used to cluster and analyze the user attributes, and several types of user attributes were obtained. The prototype was used as the dimension frame to construct the portrait, and the sample users were investigated through the offline qualitative research method to explore the deep-level demand motivation. Eventually, several types of user portrait prototypes were built. Taking the mobile phone camera photo behavior research project of young Chinese users as an example, the user''s photo-related comment information is obtained through the public data on online data platform and e-commerce platform, and text mining and analysis is carried out. The analyzed data is extracted, clustered and qualitatively studied to obtain three kinds of user portrait models of different photo types.