Recommendation engine with profile analysis

• US 9,754,306 B2
• Filed: 05/16/2014
• Issued: 09/05/2017
• Est. Priority Date: 03/03/2014
• Status: Active Grant

• Current Assignee: Invent.ly LLC
• Original Assignee: Invent.ly LLC
• Inventors: Brown, Stephen J.
• Primary Examiner(s): Rudy, Andrew Joseph
• Application Number:
  US14/280,547
• Publication Number:
  US 20150248721A1
• Time in Patent Office: 1,208 Days
• Field of Search: 725 46, 725 47
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04N 21/4532 involving end-user characteristics, e.g. viewer profile, preferences monitoring of user activities for profile generation for accessing a video database G06F16/739; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18
• CPC Class Codes: H04N 21/47 End-user applications

Abstract

A computer implemented system and method includes a recommendation engine that provides accurate recommendations based on an accurate analysis of the tastes and preferences of a user. The recommendation engine takes into consideration the information corresponding to the tastes and preferences of the users using information gathered from social networking profiles of the user as well as the reviews previously provided by the user. The recommendation engine collaborates this information with the review related information obtained from the reviewers in order to ascertain the recommendations that would match the preferences and tastes of the user.

Claims

1. A computer implemented system for generating recommendations corresponding to a plurality of items, said system comprising:

a prompter accessible to a first user, wherein said prompter prompts said first user to input social network profile information of said first user, and wherein said prompter prompts said first user to input at least one item name for an item and assign a user score to said item;

a first search module operatively connected to said prompter, wherein said first search module searches for and elicits a first group of at least one review score generated by respective reviewers and corresponding to said item reviewed by said first user, and wherein said first search module searches for and elicits a second group of at least one review score generated only by at least one reviewer who generated said first group of at least one review score;

a processor operatively connected to said first search module, wherein said processor calculates a plurality of weight scores based on said user score and said first group of at least one review score, wherein said processor assigns said plurality of weight scores to said respective reviewers, and calculates an average weight corresponding to each of said respective reviewers, and wherein said processor calculates probability scores corresponding to each item linked to said second group of at least one review score;

a recommendation engine operatively connected to said processor, said recommendation engine comprising;

a generator that generates a list of recommendations comprising said at least one item name linked to said second group of at least one review score, wherein said list is generated based on at least said probability scores corresponding to each of item linked to said second group of at least one review score, wherein said generator provides said first user with access to said list of recommendations for review of items thereof, and wherein said generator iteratively regenerates said list of recommendations to include only names of said items yet to be reviewed by said first user and linked to said second group of at least one review score; and

an updating module operatively connected said generator, wherein said updating module updates at least said social network profile information of said first user with information corresponding to items reviewed by said first user;

a clustering module operatively connected to said prompter, wherein said clustering module has access to respective social networking profiles of a plurality of users, said respective social networking profiles being updated by said updating module, and wherein said clustering module segregates said respective social networking profiles of said plurality of users into a plurality of clusters based on at least a taste and preference elicited from said respective social networking profiles; and

a second search module operatively connected and cooperating with said clustering module for accessing each of said plurality of clusters, wherein said second search module searches for and identifies said at least one item name and corresponding user scores available in at least one cluster common to at least one of said plurality of users and said first user, and wherein said search module instructs said generator to iteratively regenerate said list of recommendations to include names of items identified by said second search module.

2. The system of claim 1, wherein said prompter prompts said first user to input social network information of said first user including a user name and password for a respective social networking profile.

3. The system of claim 1, wherein said processor calculates said plurality of weight scores using a function of squares of difference between each of said at least one review score in said first group and said user score.

4. The system of claim 1, wherein said prompter instructs said generator to iteratively regenerate said list of recommendations by eliminating from said list of recommendations at least one item that is selected for review by said first user.

5. The system of claim 1, wherein said clustering module groups said respective social networking profiles of said plurality of users into said plurality of clusters based on interests of said first user and demographic information corresponding to said first user.

6. The system of claim 1, wherein said probability scores calculated by said processor are indicative of a probability of said first user preferring each said item present in said second group, and wherein said probability scores are functions of respective said at least one review score present in said second group and an average weight for said respective reviewers.