Activity and action prediction for an emphatic space using context

• Main Authors: Bautista, Nikka Jennifer G., Cua, Manuel M., Jr., Aureus, Jed J., Urquiola, Marc Angelo B.
• Format: text
• Language: English
• Published: Animo Repository 2010
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/11956
• Institution: De La Salle University

Emerging technologies call for an improvement in the way humans and computers interact. The construction of a highly responsive, emphatic environment that facilitates automatic monitoring and supporting of life occupants calls for the need for recognition and prediction of the occupant’s activities. Recent studies regarding the topics involve the use of methods that range from using First-Order Predicate Logic (FOPL) to instance-based learning to statistical learning algprthms like Hidden Markov Models (HMMs), Naïve Bayes Networks, Bayesian Networks and C4.5 which have produced relatively accurate results with regards to the recognition of human activity but very little effort has been allocated for activity prediction. Majority of the researches focus on activity recognition because the automatic and unobtrusive identification of human activity is a fundamental and based on popular opinion, one of the most challenging goals in context-aware computing. Lingering issues with regards to activity recognition include the proper representation of actions for learning and understanding robust activity models, effective methods to automatically build a model of the user’s activities of daily living (ADL), real-time activity recognition and prediction. In this research, the proponents explored the use of Javabugs, a tree-based modeling algorithm, for the clustering of actions in building a model for the prediction of a user’s actions within the emphatic.