SURGE aims to provide recommendations based on the CoSKQ algorithm.
CoSKQ intents to retrieve best-cost subsets from a pool of possibilities given a query. The query and every element in the set of data consists of physical coordinates and a set of keywords. In order to calculate the costs of a subset, CoSKQ calculates the physical distance between the query and the subset and also the physical distance between the entries of the subset. These calculations are only carried out if the combined keywords of the subset cover the keywords of the query.
Plain CoSKQ is not suitable for recommender systems, as it requires complete matches between the subset's keywords and the query's keywords. Therefore, SURGE calculates these best-cost subsets based on their physical distance to the query, the distance between the subset entries and the similarity between the query and subset keywords. However, SURGE does not strictly exclude subsets because of the keyword coverage.
