Space reduction for contextual transaction trust computation in e-commerce and e-service environments

In the literature, Contextual Transaction Trust computation (termed as CTT computation) is considered an effective approach to evaluate the trustworthiness of a seller. Specifically, it computes a seller's reputation profile to indicate his/her dynamic trustworthiness in different product categories, price ranges, time periods, and any necessary combination of them. Then, in order to promptly answer a buyer's requests on the results of CTT computation, CMK-tree has been designed to appropriately index the precomputed aggregation results over large-scale ratings and transaction data. Nevertheless, CMK-tree requires additional storage space. In practice, a seller usually has a large volume of transactions. Moreover, with significant increase of historical transaction data (e.g., Over one or two years), the size of storage space consumed by CMK-tree will become much larger. In reducing storage space consumption for CTT computation, the aggregation results that are generated based on the ratings and transaction data from remote history, e.g., '12 months ago' can be deleted, as the ratings from remote history are less important for evaluating a seller's recent behavior. However, to achieve nearly linear and robust query performance, the deletion operations in the CMK-tree become complicated. In this paper, we propose three deletion strategies for CTT computation based on CMK-tree. With our proposed deletion strategies, the additional storage space consumption can be restricted to a limited range, which offers great benefit to trust management with millions of sellers. Finally, we have conducted experiments to illustrate both advantages and disadvantages of the proposed deletion strategies.