STAGE:

The internet transactional environment is so vast that it’s nearly impossible to have a monitoring authority to cross-check the fraudulent practices. For example, a seller can list a product of well-known brand for selling, but delivers a fake or cheaper version of the same while the buyers had already paid for it and but now has no way to raise his concern. To minimise such incidents it’s important that buyer must have enough information about the seller, and his reputation. Several statistical and mathematical solutions have been developed to make users aware of such practices, but those solutions have their own shortcomings. For example, beta reputation systems can compute the aggregate trust rating for a seller, but it requires several iterations to filter out the ratings which could be intentional and unfair. And even after employing those iterations, there is no conformity about the correctness and accuracy of the results. Moreover, these methods require several truncations to be completed to compute the results. In the absence of such information, these methods can be attacked by fraudulent techniques like: 1.) “bootstrapping”, and 2.) “whitewashing”.

Bootstrapping occurs when a seller has just entered the online market and there is no evidence that he could be trusted or not. The statistical reputations systems, in such cases, assume an average priori trust value indicating the seller could or couldn’t be trustworthy. If priori trust value is set above average, the sellers who were already detected using fraudulent means to sell the products, can easily break the system and create new identity to enter the market. This practice is called whitewashing. If priori trust value is set below average, new sellers who could be genuine will find it difficult to sell their products.

In recent years, there have been great studies regarding