Recommendation system with multiple integrated recommenders

• US 8,260,787 B2
• Filed: 06/29/2007
• Issued: 09/04/2012
• Est. Priority Date: 06/29/2007
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Lam, Shing Yan, Chakrabarti, Kushal, Ionkov, George M., Kim, Sung H., Witt, Brett W.
• Primary Examiner(s): Alam, Hosain
• Assistant Examiner(s): Ahn, Sangwoo
• Application Number:
  US11/771,914
• Publication Number:
  US 20090006398A1
• Time in Patent Office: 1,894 Days
• Field of Search: None
• US Class Current: 707/748
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages

Abstract

A recommendations system is provided in various embodiments for selecting items to recommend to a user. The system includes a recommendation engine with a plurality of recommenders, and each recommender identifies a different type of reason for recommending items. In one embodiment, each recommender retrieves item preference data and generates candidate recommendations responsive to a subset of that data. The recommenders also score the candidate recommendations. In certain embodiments, a normalization engine normalizes the scores of the candidate recommendations provided by each recommender. A candidate selector selects at least a portion of the candidate recommendations based on the normalized scores to provide as recommendations to the user. The candidate selector also outputs the recommendations with associated reasons for recommending the items.

Claims

1. A computer-implemented method of normalizing item recommendation scores, the method comprising:

by a computer system comprising computer hardware;

assigning scores to candidate recommendations with first and second recommenders, the first recommender configured to assign first scores to a first portion of the candidate recommendations, the second recommender configured to assign second scores to a second portion of the candidate recommendations using a different scoring scale than the first recommender;

outputting the scores from each of the recommenders to a normalization engine;

for each recommender, normalizing the scores output by the recommender with the normalization engine by at least;

combining the scores for at least some of the candidate recommendations to generate a combined score, said combining comprising computing a moving average of at least some of the scores, and

calculating normalized scores based at least in part on the combined score and the scores for at least some of the candidate recommendations; and

using the normalized scores to select at least a portion of the candidate recommendations to recommend to the target user.

2. The method of claim 1, wherein said translating at least some of the scores comprises translating at least some of the scores by a difference between a minimum score and the desired normalized minimum score in response to determining that the minimum score differs from the desired normalized minimum score.

3. The method of claim 1, further comprising adjusting the number of scores combined in the combined score to reduce the impact of outliers on the normalized scores.

4. The method of claim 1, wherein combining the scores for at least some of the candidate recommendations comprises summing the scores.

5. The method of claim 1, further comprising assigning weights to candidate recommendations received from the first and second recommenders.

6. The method of claim 1, wherein calculating normalized scores based at least in part on the combined score and the scores for at least some of the candidate recommendations comprises dividing each candidate recommendation score by the combined score.

7. The method of claim 1, wherein the moving average comprises an exponential moving average.

8. A system for normalizing item recommendation scores, the system comprising:

a system comprising computer hardware, the system programmed to implement;

a plurality of recommendation algorithms, each of the recommendation algorithms configured to;

generate candidate recommendations, assign scores to the candidate recommendations, and to output the candidate recommendations and associated scores, wherein at least one of the recommendation algorithms is configured to use a different scoring scale for scoring candidate recommendations than another one of the recommendation algorithms;

a normalization engine operative to normalize the scores output by the plurality of recommendation algorithms, the normalization engine configured to at least;

combine the scores for at least some of the candidate recommendations to generate a combined score by at least computing a moving average of at least some of the scores, and

calculate normalized scores based at least in part on the combined score and of the scores for at least some of the candidate recommendations; and

a candidate selector configured to select at least a portion of the candidate recommendations to recommend to a target user based at least in part on the normalized scores.

9. The system of claim 8, wherein the normalization engine is further operative to combine the scores for at least some of the candidate recommendations by at least computing a moving average of a subset of the scores.

10. The system of claim 9, wherein the normalization engine is further operative to calculate normalized scores based at least in part on the combined score and of the scores for at least some of the candidate recommendations by at least dividing each candidate recommendation score by the combined score.