The creation of a daily hierarchy and a dynamic clustering method, you can build a recommendation service that is specific to the clients’ context based on the multidimensional recommendation model. [1] Clients’ profile data (name, age, gender, single/married, location, and some registered information). Clients are dynamically clustered based on contextual information and getting clients to rate the recommendations and help to modify the accuracy of suggestions using a hidden Markov chain model.[2]
My approach is to use time, location, documents as the contextual information dimensions and then refine the process as more clients get on the system (~10,000 clients is my target)
Refine Recommendation
- Using History of User’s Actions [3]
- Using Commonsense algorithm[4]
- Using Case-Based Reasoning algorithm [5]
- Using Sensor Data [6]
[1] Wang, C.-Y., Wu, Y.-H. and Chou, S.-C. T. (2008). Toward a ubiquitous personalized daily-life activity recommendation service with contextual information: A services science perspective, HICSS ’08: Proceedings of the Proceedings of the 41st Annual Hawaii International Conference on System Sciences, IEEE Computer Society, Washington, DC, USA, p. 107.
[2] Intelligent recommendation algorithm based on hidden Markov chain model.
[3] Naeem, U. and Bigham, J. (2007). A comparison of two hidden markov approaches to task identification in the home environment, Pervasive Computing and Applications, 2007. ICPCA 2007. 2nd International Conference on, pp. 383–388.
[4] Lieberman, H. and Espinosa, J. (2007). A goal-oriented interface to consumer electronics using planning and commonsense reasoning, Know.-Based Syst. 20(6): 592–606.
[5] Ni, H., Zhou, X., Zhang, D., Miao, K. and Fu, Y. (2009). Towards a task supporting system with cbr approach in smart home, ICOST ’09: Proceedings of the 7th International Conference on Smart Homes and Health Telematics, Springer-Verlag, Berlin, Heidelberg, pp. 141–149.
[6] Bouarfa, L., Jonker, P. and Dankelman, J. (2010). Discovery of high-level tasks in the operating room, Journal of Biomedical Informatics .