In this research article, we present our work on building computational prediction models to dynamically predict users purchase behaviours by implementing Hidden Markov Models (HMM). The models can be used by decision makers in a company to develop a strategy (e.g. marketing, products development) based on the prediction results. We evaluate the model using our datasets of Facebook. We collected the data by utilising Facebook API.
Furthermore, we implemented a Hidden Markov Model (HMM) algorithm to the datasets to provide a dynamic prediction of customers purchase behaviours over time. In the preliminary evaluation, we implemented our model to the datasets with t=2. In our datasets, we found that the category, electronics, was the most favourite topic to discuss, share and
like regarding electronics. Interestingly, we found that a positive direction for its trend appeared in the second run of the model.
Lasmy, Chowanda A, Herman D. R. T and Notoatmodjo B. (2016). Towards Automatic Customer Purchase Behaviours Prediction through a Social Media Lens Using the Hidden Markov Model. Pertanika Journal of Social Science and Humanities, 24, 169-176.
Customer purchase behaviour, Hidden Markov Model, Facebook datasets, strategic management, computational prediction mode