Maintaining preference networks that adapt to changing preferences
University of New Brunswick
In recent years, online transaction systems have used automated decision making or automated negotiation technology to interact with their customers by suggesting other products or services that might have high utility for those customers. However, decision making can be more difficult with an enormous amount of information, not only for the human but also for the automated decision making processes. Without a proper user preference model, search results are often inaccurate, and thus may result in many pages of output for users to examine. Although most user preference elicitation models have been developed based on the assumption that user preferences are stable, user preferences may change in the long term and may evolve with experience, resulting in dynamic preferences. Therefore, in this thesis, we provide a model called a Dynamic COP-net that is maintained using an approach that does not require the entire preference graph to be rebuilt when a previously-learned preference is changed, with efficient algorithms to add new preferences and to delete existing preferences. A Dynamic COP-net can maintain the preference model without rebuilding the entire preference model whenever there is a new preference added or deleted, within O(n²) time. Dynamic COP-nets are shown to outperform existing algorithms for insertion, especially for large numbers of attributes and for dense graphs. They do have some shortcomings in the case of deletion, but only when there is a small number of attributes or when the graph is particularly dense.