US Patent No. 9449322 - Prepared by Attorney David Tran for eBay and filed by Schwegman (SLWK)
Brief description: For some example embodiments, the desirability scores of the terms may depend on how the terms are used in the perspective of demand as related to search queries and in the perspective of supply as related to items. As for the demand, the data analysis application(s) 207 may receive search queries initiated by users via a search interface. The data analysis application(s) 207 may extract terms (or keywords) from each of the search queries. The extracted terms may then be added to a set of stored terms to generate updated stored terms. For each term, a demand percentage is calculated or recalculated based on the updated stored terms. As for the supply, a total occurrence of unique terms in the title information of stored items may be determined. A percentage occurrence of each term is then determined based on the result of the total occurrence. For example, if there were 1000 items in the result, and the term “charger” occurs in 900 items, then the term “charger” is associated with a supply percentage of 90%. When the term “player” occurs in 50 items, then its associated supply percentage is 5%. A formula may be used to determine the desirability score of a particular term based at least on its demand percentage and its supply percentage. The desirability score of the terms may be stored in the database 108. It may be noted that a desirability score may be associated with a positive number (desirable) or a negative number (not desirable). For example, when a user specifies a search query that includes the phrase “Ipod Nano”, corresponding terms and their desirability scores may include “mp3 90%”, “player 85%”, “transmitter −40%”, “case −80%”, and so on. For some example embodiments, the suggestion application(s) 209 may include operations that suggest terms based on their desirability scores and/or based on co-occurring terms. The suggestion application(s) 208 may extract one or more terms from the title information of the item being listed. Based on the extracted terms, co-occurring terms may be determined as suggested to the user. The co-occurring terms and/or the terms with the high desirability scores may be suggested as candidates that may be used to modify the title information. It may be noted that the suggestion application(s) 209 may extract the one or more terms from other section (e.g. item description information, etc.) of the listing besides the title information. It may also be noted that the suggestion application(s) 209 may extract concepts from the information associated with the item being listed. A concept may include two or more terms.