DISCOVERING SPAM MERCHANTS USING PRODUCT FEED SIMILARITY

摘要

Discovering spam merchants using product feed comparison comprises a merchant signature computed by a comparison system using the offer data submitted by a merchant. The merchant submits a file to the comparison system for use on a product comparison website. The comparison system processes the file and calculates a merchant signature by parsing a specific feature common to all offers. The comparison system assigns a data value to each of the specific features included in each of the offers and generates a string of assigned data values. The comparison system executes a hash algorithm for the string of data values to obtain the merchant signature. The merchant signature is compared with the signatures of known spam merchants and if the merchant signature is within a predefined threshold of known spam merchant signatures, the merchant is rejected and the signature is marked as a spam merchant.