Churn prediction in new users of Yahoo! answers

Abstract

One of the important targets of community-based question answering (CQA) services, such as Yahoo! Answers, Quora and Baidu Zhidao, is to maintain and even increase the number of active answerers, that is the users who provide answers to open questions. The reasoning is that they are the engine behind satisfied askers, which is the overall goal behind CQA. Yet, this task is not an easy one. Indeed, our empirical observation shows that many users provide just one or two answers and then leave.In this work we try to detect answerers that are about to quit, a task known as churn prediction, but unlike prior work, we focus on new users. To address the task of churn prediction in new users, we extract a variety of features to model the behavior of Yahoo! Answers users over the first week of their activity, including personal information, rate of activity, and social interaction with other users. Several classifiers trained on the data show that there is a statistically significant signal for discriminating between users who are likely to churn and those who are not. A detailed feature analysis shows that the two most important signals are the total number of answers given by the user, closely related to the motivation of the user, and attributes related to the amount of recognition given to the user, measured in counts of best answers, thumbs up and positive responses by the asker.

A Churn Prediction Iterative Model in Social Network Based on Relationship Strength

“…Zhu Y et al [14]studied user activity levels in social network, Long X et al [17] researched user relationship in social network.User churn model: there are many domestic and foreign scholars using data mining method to build user churn model, including classification model, the regression model, the clustering model, time series analysis. Studies such as Dror G et al [18] tried a variety of models in the churn prediction, including Logistic regression, SVM, KNN. Studies such as Oentaryo R J et al [9] used SVM and based on SVM improved ICA model.…”

Modelling User Collaboration in Social Networks Using Edits and Comments

“…Badges are awarded in several categories including question, answer, participation, tag and moderation badges. We only considered badges in the answer category because we are interested in collaboration in answer posts since quality answers is one of the factors that keeps a Q&A social network active [3].…”

“…In Q&A social networks, collaboration between authors results in better answers [2] and quality answers result in an active Q&A network [3]. Hence, the importance of studying collaboration among users who answer questions cannot be overemphasized.…”

Modeling Attrition in Organizations from Email Communication

“…Thus identifying people who are likely to churn early would allow the platform to investigate such users and make changes (or add features) to rectify such a situation. In [23], [24], the authors try to identify user features that can lead to churn in social networks. They then use these features to train classifiers for churn prediction achieving fair accuracy.…”