Transaction Similarity-Based Contextual Trust Evaluation in E-Commerce and E-Service Environments

2013 IEEE 20th International Conference on Web Services

Orgun

2013

Self Cite

Abstract-Online trading takes place in a very complex environment full of uncertainty in which deceitful service providers or sellers may strategically change their behaviors to maximize their profits. The proliferation of deception cases makes it essential to model the dynamics of a service provider and predict the trustworthiness of the service provider in transactions. Recently, probabilistic trust models have been used to assist decision making in computing environments. Although the typical Hidden Markov Model (HMM) has been used to model a provider's behavior dynamics, existing approaches focus only on the outcomes or ignore the hidden characteristics of the HMM model. In this paper, we model the dynamic trust of service providers concerning a forthcoming transaction in light of as much information as we can consider, including the static features, such as the provider's reputation and item price, and the dynamic features, such as the latest profile changes of a service provider and price changes. Based on a service provider's historical transactions, we predict the trustworthiness of the service provider in a forthcoming transaction. In addition, the Mutual Information theories and the Principle Component Analysis method are leveraged to eliminate redundant information and combine essential features to form lower dimensional feature vectors. Furthermore, by adopting Vector Quantization techniques, we apply the discrete HMM in a more powerful way, in which all the features extracted from both contextual information and the rating of each transaction are treated as observations of HMM. We evaluate our approach empirically in order to study its performance. The experiment results illustrate that our approach significantly outperforms the state-of-theart probabilistic trust methods in accuracy in the cases with complex changes.

Section: Introductionmentioning

confidence: 99%

Modeling the Dynamic Trust of Online Service Providers Using HMM

Zheng

2013 IEEE 20th International Conference on Web Services

Orgun

2013

Self Cite

“…In our previous work [22], [24], we have analyzed and summarized the following reasons in most existing trust evaluation models why they cannot identify value imbalance in transactions.…”

Section: Introductionmentioning

confidence: 99%

A Novel Model for Contextual Transaction Trust Computation with Fixed Storage Space in E-Commerce and E-Service Environments

Zhang

2013 IEEE International Conference on Services Computing

2013

Self Cite

Abstract-In e-commerce and e-service environments, transaction context is important when evaluating the trust level of a seller or a service provider in a forthcoming transaction. However, most existing trust evaluation models compute a single value to reflect the general trust level of a seller without taking any transaction context into account. In the literature, a trust vector approach has been proposed to resolve the above problem. In particular, the trust vector contains different sets of trust values (termed as CTT values) so as to outline a seller's reputation profile. As a result, buyers can identify the potential risk existing in a forthcoming transaction (e.g., value imbalance, i.e. a malicious seller may build up a high level of trust by selling cheap products and then deceive buyers by inducing them to purchase more expensive products) and thus avoid monetary losses.In computing CTT values, some approaches are proposed that store the precomputed aggregation results over large-scale ratings and transaction data of a seller, so as to deliver prompt responses to a buyer's query. Though these approaches allocate relatively small space to each seller for storing the aggregation results, if applied in a system with millions of sellers, space consumption will be intolerable. In this paper, we propose a novel model for CTT computation with fixed storage space, which provides a trade-off between aggregation detail and storage space. It is particular suitable for CTT computation where a request is regarding a seller's trust in recent time period, e.g., the latest six months, rather than six months plus one day. We have conducted experiments on both an eBay dataset and a synthetic dataset to illustrate its good efficiency in responding to buyers' CTT queries.

“…Hence, the transaction item similarity S T I between two products p and p should be a monotonically decreasing function with respect to the depth d of the deepest common ancestor of them. In our earlier work [23], we have proposed a formula to calculate the similarity within a hierarchical structure of product category. Due to space constraints, this paper does not include detailed explanation.…”

Section: A Similarity Comparison Of Transaction Itemsmentioning

confidence: 99%

“…Caballero et al [3] defined a formula using task key values to calculate the similarity between two tasks in order to evaluate the trust level of different tasks. In e-commerce environments, Zhang et al [23] introduced the concept of similarity between past transactions and a forthcoming transaction when evaluating transaction trust, then proposed some formulae to calculate transaction context similarity to infer the trust value of a forthcoming transaction.…”

Section: B Context-aware Trust Evaluationmentioning

confidence: 99%

A trust vector approach to transaction context-aware trust evaluation in e-commerce and e-service environments

Zhang

2012 Fifth IEEE International Conference on Service-Oriented Computing and Applications (SOCA)

Zhang

2012

Self Cite

Abstract-At some e-commerce websites (such as eBay), a trust value of a seller is computed based on the ratings of past transactions given by buyers, which can only reflect the general or global trust level of a seller without any transaction context information taken into account. As a result, a buyer may be easily deceived by a malicious seller in a forthcoming transaction. For example, with the notorious value imbalance problem, a malicious seller can build up a high trust level by selling cheap products and then start to deceive buyers in selling expensive products. In this paper, we first model all contextual transaction factors that reflect the nature of transactions, and thus influence the evaluation of transaction trust. In addition, instead of providing a single trust value, we propose a trust vector approach that takes into account the contextual factors in transactions. Our model systematically categorize these factors into service aspect and transaction aspect. In particular, the computation of the elements in this trust vector is associated with both the context of past transactions and the context of a forthcoming transaction, so as to comprehensively indicate the trust level of a seller for the forthcoming transaction. The computed trust vector can be taken as the reputation profile of the seller. Empirical studies illustrate that it is important and necessary to introduce contextual transaction factors in evaluating the trust level of sellers objectively.