社交网络群中用户活跃度分析与预测

On Prediction and Analysis of Social Network Group User Activity

对于社交网络中不同的群组,由于用户属性(性别、年龄等)、群类别、群成员之间关系等因素的影响,其活跃度各不相同.本文首先从社交网络用户数据中提取人口信息、群的类别、社交关系、群用户黏性(分享消息数、图片数)等特征,然后利用logistic回归、支持向量机、BP神经网络等机器学习算法对不同群中用户的活跃度进行预测.结果表明,BP神经网络针对社交网络群中用户活跃度分类判断时具有较高的预测性能,社交关系特征对群用户活跃性具有重要影响.

For different groups in a social network,their activities are frequently influenced by a variety of factors such as user's attributes(i.e.,gender,age),group classes,social relationships between group members and so on.In order to model and analyze the activities of different groups in this paper,several features which may influence the activity of a group have first been extracted from the historical data generated from a social network,such as census information,social relationships,group class,user stickiness(the number of the shared photos and information)and so on.Then,based on the extracted features,the activity of a group is predicted using logistic regression model,support vector machine and BP neural network.The results show that BP neural network has high performance on classifying group users activity,and social relationships have a major impact on the activity of a group.