A Performance Evaluation of Preference Evaluation Techniques in Real High Dimensional Database

Preference query has received high interest due to its great benefits over various types of database applications. This type of query provides more flexible query operators that retrieve data items which are not dominated by the other data items in all attributes (dimensions). Many preference techniques for preference queries have been introduced including top-k, skyline, multi-objective skyline, top-k dominating, k-dominance, ranked skyline, and k-frequency. All of these preference techniques aimed at finding the “best” result that meets the user preferences. This paper aims at evaluating the performance of the five well-known preference evaluation techniques, namely: top-k, skyline, top-k dominating, k-dominance and k-frequency; in a real database application when high number of dimensions is the main concern. To achieve this, a recipe searching application with maximum number of 60 dimensions has been developed which assists users to identify the most desired recipes that fulfill their preferences. Several analyses have been carried out, where execution time is the main measurement used to evaluate each preference technique