A system and method for determining vendor identity based on electronic documents analysis. The method includes receiving, by a computing device, a first electronic document. The first electronic document includes a plurality of textual items. The method also includes extracting, using the computing device, at least a first sequence of textual items from the plurality of textual items, determining a first approximate identity of a vendor entity that issued the first electronic document by comparing the first sequence of textual items to at least a second sequence of textual items that is associated with at least a second electronic document of a plurality of second electronic documents that was previously associated with a unique vendor entity, and extracting at least a first multiset of textual items from the plurality of textual items of the first electronic document.