Purpose: Though online communities offer unprecedented opportunities to involve people in knowledge sharing, the reasons why users would like to participate in those activities in online communities have still been under-explored. In this research, the authors aim to use the value co-creation theory to build and test a continuance usage model, which focuses more on experiential values resulted from the knowledge sharing behavior. Design/methodology/approach: An integrative research model is built to investigate how knowledge sharing behavior affects users’ co-creation value and then drives their continuance usage in online communities. Online survey data collected from 239 Sina Microblog users in China are utilized to validate the proposed model and hypotheses. Findings: Empirical results indicate that the knowledge sharing behavior helps improve users’ co-creation value, including customer learning value, social integrative value and hedonic value. This co-creation value can subsequently affect users’ future participation intention in online communities. Originality/value: This paper seeks to fill the research gap by examining customers’ motivations or perceptions underlying their knowledge sharing behavior at the usage-stage, instead of the pre-usage stage mainly concentrated on by previous studies. The managerial implications can be utilized for policy making to encourage customers’ participation and operate a better online community.

In this study, we extend the traditional monetary model and the random walk model with Markov-switching method and propose two new forecasting models called the Markov switching monetary model (MSMM) and Markov switching random walk model (MSRW). Then, we evaluate the forecasting ability of these two new mixed models, MSMM and MSRW, and compare their performance with the traditional pure monetary model and pure Random walk model based on Mean Squared Forecast Error and Mean Absolute Forecast Error. The results show that the two hybrid models significantly improve the forecasting ability compared with the two traditional models in most scenarios. Moreover, we reexamine the role of data frequency in determining the number of regimes and in affecting the accuracy of forecast evaluation with different data frequency.