Methods and systems for simulating a search to generate an optimized scoring function

• US 10,140,339 B2
• Filed: 01/26/2010
• Issued: 11/27/2018
• Est. Priority Date: 01/26/2010
• Status: Active Grant

• Current Assignee: PayPal, Inc. (PayPal Holdings, Inc.)
• Original Assignee: PayPal, Inc. (PayPal Holdings, Inc.)
• Inventors: El-Charif, Rami, Ghatare, Sanjay Pundlkrao, Chen, Steven, Dumon, Olivier G., Rehman, Muhammad Faisal, Song, Guanglie
• Primary Examiner(s): Gorney, Boris
• Assistant Examiner(s): Roland, Griselle C
• Application Number:
  US12/694,191
• Publication Number:
  US 20110184883A1
• Time in Patent Office: 3,227 Days
• Field of Search: 707748, 707754, 707723
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 10/067 Enterprise or organisation modelling
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Methods and systems for deriving a scoring function, for use in ordering a set of search results, are described. Consistent with some embodiments, a simulation platform includes an offline simulator module that receives search result sets for search queries that result in the conclusion of a transaction. The result set is then processed by the offline simulation platform to derive a set of weighting factors for use with one or more parameters in a parametric scoring function. The weighting factors are determined, for example, by specifying an equation setting one or more constraints and then solving the equation for the weighting factors.

Claims

1. A computer-implemented method comprising:

with a processor-based data importing module, importing data representing a plurality of search result sets resulting from a plurality of executed search queries, each search result set including a plurality of item listings satisfying a search query from the plurality of executed search queries and a listing slot identifier identifying a position within a search results page of a particular item listing that has resulted in a transaction being concluded, the listing slot identifier determined utilizing a production scoring function specified with one or more parameters that are weighted with a first set of one or more weighting factors;

with a processor-based parameter optimization module, processing the plurality of search result sets to derive a new scoring function, the new scoring function having a second set of one or more weighting factors for one or more parameters, the second set of one or more weighting factors selected to satisfy a set of constraints, the second set of one or more weighting factors being different than the first set of one or more weighting factors, wherein the second set of one or more weighting factors are selected to maximize an average rank shift metric for the plurality of search result sets, the average rank shift metric determined by comparing the listing slot identifier as derived by the production scoring function with a listing slot identifier as derived by a new scoring function for each item listing that has resulted in a transaction being concluded for each search result set of the plurality of search result sets; and

with a processor-based scoring function evaluation module, evaluating the performance of the new scoring function relative to a production scoring function.

2. The computer-implemented method of claim 1, further comprising:

filtering the plurality of search result sets resulting from the plurality of executed search queries to select search result sets having a predetermined number of item listings assigned to a particular category of item listings, wherein processing the plurality of search result sets to derive a new scoring function includes processing the search result sets having the predetermined number of item listings assigned to the particular category.

3. The computer-implemented method of claim 2, wherein the one or more parameters are selected based on a particular category of item listings and the new scoring function is for use with search queries having a predetermined number of search results that are item listings assigned to the particular category.

4. The computer-implemented method of claim 1, wherein a slot identifier is a number derived by the production scoring function and assigned to an item listing satisfying a search query, the number used in ordering the item listings that satisfying the search query when presenting the item listings in a search results page.

5. The computer-implemented method of claim 1, wherein the new scoring function is specified with the same parameters as the production scoring function, but different weighting factors than the production scoring function.

6. The computer-implemented method of claim 1, wherein the new scoring function is specified with parameters that differ from the production scoring function.

7. A data processing system comprising:

at least one processor; and

a machine-readable medium in communication with the at least one processor, the machine readable medium storing a simulation platform module that is executable by the at least one processor, the simulation platform module being executed by the at least one processor to cause operations to be performed, the operations comprising;

importing data representing a plurality of search result sets resulting from a plurality of executed search queries, each search result set including a plurality of item listings satisfying a search query from the plurality of executed search queries and a listing slot identifier identifying a position within a search results page of a particular item listing that has resulted in a transaction being concluded, the listing slot identifier determined utilizing a production scoring function specified with one or more parameters that are weighted with a first set of one or more weighting factors;

processing the plurality of search result sets to derive a new scoring function, the new scoring function having a second set of one or more weighting factors for one or more parameters, the second set of one or more weighting factors selected to satisfy a set of constraints, the second set of one or more weighting factors being different than the first set of one or more weighting factors, wherein the second set of one or more weighting factors are selected to maximize an average rank shift metric for the plurality of search result sets, the average rank shift metric determined by comparing the listing slot identifier as derived by the production scoring function with a listing slot identifier as derived by a new scoring function for each item listing that has resulted in a transaction being concluded for each search result set of the plurality of search result sets; and

evaluating the performance of the new scoring function relative to a production scoring function.

8. The data processing system of claim 7, wherein the one or more parameters are selected based on a particular category of item listings and the new scoring function is for use with search queries having a predetermined number of search results that are item listings assigned to the particular category.

9. The data processing system of claim 7, wherein a slot identifier is a number derived by the production scoring function and assigned to an item listing satisfying a search query, the number used in ordering the item listings that satisfying the search query when presenting the item listings in a search results page.

10. The data processing system of claim 7, wherein the new scoring function is specified with the same parameters as the production scoring function, but different weighting factors than the production scoring function.

11. The data processing system of claim 7, wherein the new scoring function is specified with parameters that differ from the production scoring function.

13. The data processing system of claim 7, wherein the operations further comprise filtering the plurality of search result sets resulting from the plurality of executed search queries to select search result sets having a predetermined number of item listings assigned to a particular category of item listings, wherein processing the plurality of search result sets to derive a new scoring function includes processing the search result sets having the predetermined number of item listings assigned to the particular category.

12. An e-commerce platform, comprising:

at least one processor in communication with a machine-readable storage medium storing an on-line trading application module that is executable by the at least one processor, the on-line trading application module being executed by the at least one processor to utilize a production scoring function to order sets of item listings that satisfy search queries, the production scoring function specified with one or more parameters weighted with one or more weighting factors; and

at least one processor in communication with a machine-readable storage medium storing a simulator module that is executable by the at least one processor, the simulator module to perform operations comprising;

receiving data representing sets of ordered search results and for each set of search results a listing slot identifier identifying a listing associated with an item conversion, the search results ordered based on a listing performance score assigned to each item listing in the search results, the listing performance score derived utilizing a production scoring function having one or more parameters weighted with a first set of one or more weighting factors; and

deriving new weighting factors for a new scoring function having a second set of one or more parameters weighted with the new weighting factors, the new scoring function and the second set of one or more weighting factors derived based on one or more specified constraints, the second set of one or more weighting factors being different than the first set of one or more weighting factors, wherein the second set of one or more weighting factors are selected to maximize an average rank shift metric for the set of search results, the average rank shift metric determined by comparing the listing slot identifier as derived by the production scoring function with a listing slot identifier as derived by a new scoring function for each item listing that has resulted in a transaction being concluded for each search result of the set of search results.