Context-aware transaction trust computation in e-commerce environments

In e-commerce environments, the trustworthiness of a seller is very important to potential buyers, especially when the seller is not known to them. Most existing trust evaluation models compute a single value to reflect the general trustworthiness of a seller without taking any transaction context information into account. With such a result as the indication of reputation, a buyer may be easily deceived by a malicious seller in a transaction where the notorious value imbalance problem is involved, namely, a malicious seller accumulates a high level reputation by selling cheap products then deceives buyers by inducing them to purchase more expensive products. This thesis aims to systematically investigate some key and open challenging research problems in context-aware transaction trust computation in e-commerce environments. In general, it includes our work from the following two aspects. The first aspect is the trust vector based approach to context-aware transaction trust evaluation. In contrast to most existing trust management models that compute a single trust value, a trust vector is first presented consisting of three major values for Contextual Transaction Trust (CTT). In the computation of CTT values, three identified important context dimensions, including Product Category, Transaction Amount and Transaction Time, are taken into account. In the meantime, the computation of each CTT value is based on both past transactions and the forthcoming transaction. In particular, with different parameters specified by a buyer regarding context dimensions, different sets of CTT values can be calculated. As a result, all these trust values can outline the reputation profile of a seller that indicates his/her dynamic trustworthiness in different products, product categories, price ranges, time periods, and any necessary combination of them. We term this new model as ReputationPro. Nevertheless, in ReputationPro, the computation of reputation profile requires new data structures for appropriately indexing the pre-computation of aggregates over large-scale ratings and transaction data in three context dimensions, as well as novel algorithms for promptly answering buyers’ CTT requests. In addition, storing pre-computed aggregation results consumes a large volume of space, particularly for a system with millions of sellers. Therefore, reducing storage space for aggregation results is also a great demand. The second aspect is efficient computation of a seller’s reputation profile. Towards efficient computation of CTT values aiming at outlining a seller’s reputation profile, four index schemes have been proposed. We first extend the approaches to the two-dimensional (2D) Range Aggregate (RA) problem as the preliminary solutions for CTT computation. They are effective approaches, but have low efficiency in computing CTT values in some cases. Then, to overcome the problems in the preliminary solutions, a new disk-based index scheme and a new query algorithm are further proposed. Compared with the preliminary solutions, when answering a buyer’s CTT queries for each brand-based product category, the new index scheme has almost linear query performance. This is a significant advantage in answering queries on CTT values especially when a large number of buyers are accessing a seller’s reputation data simultaneously. In addition, several strategies are proposed for storage space reduction in CTT computation. These strategies include aggregating ratings and transaction data at different time granularity as well as deleting the index records that are generated based on the ratings and transaction data from remote history. The experiments conducted on synthetic datasets generated from eBay datasets have demonstrated that the proposed ReputationPro model can be more effectively applied to large-scale ecommerce websites in terms of efficiency and storage space consumption.