Real-Time Recommendation Confidence Thresholding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional online ordering systems lack the ability to determine real-time recommendation confidence for personalized recommendations, leading to excessive and ineffective cross-selling that wastes network bandwidth and user time.

Innovation Solution

A system that determines recommendation confidence by analyzing user profiles, anchor items, and feedback from prior recommendations to gauge the likelihood of user acceptance, transmitting recommendations only when confidence is positive and refraining from doing so when confidence is low.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the system provides recommendations for complementary items based on user cart items and purchase history, then the cross-selling effectiveness is improved, but excessive recommendations are generated that waste network bandwidth and user time

Engineering Contradiction:

Improvecross-selling effectivenessVSAvoidnetwork bandwidth waste

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The system changes the parameter of recommendation delivery by introducing a confidence threshold parameter. Recommendations are only transmitted when the calculated confidence score exceeds this threshold, transforming the recommendation system from always-on to conditional delivery based on confidence level parameters

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system implements feedback by calculating recommendation confidence scores based on user profiles, anchor items, and prior recommendation feedback. This feedback mechanism allows the system to learn from user responses and adjust recommendation confidence, preventing waste of network bandwidth on low-confidence recommendations

Inventive Principle:

Principle #23Feedback

2Productivity

If the system provides recommendations for complementary items based on user cart items and purchase history, then the cross-selling effectiveness is improved, but user time is wasted when users are slowed down by excessive recommendations

Engineering Contradiction:

Improvecross-selling effectivenessVSAvoiduser time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system applies partial action by selectively delivering only those recommendations that meet the confidence threshold, rather than delivering all possible recommendations. This partial delivery approach prevents user time waste while maintaining cross-selling effectiveness for high-confidence recommendations

Inventive Principle:

Principle #16Partial or excessive action

3Speed

If the system transmits recommendations in real-time, then the recommendation timeliness is improved, but the system complexity increases due to real-time confidence calculation requirements

Engineering Contradiction:

Improverecommendation timelinessVSAvoidsystem complexity

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

The system performs preliminary action by pre-calculating and storing user profiles, item relationships, and recommendation confidence models before real-time recommendation needs arise. This preprocessing allows rapid real-time confidence calculation without excessive system complexity during actual recommendation delivery

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary confidence calculation layer that mediates between the user profile database and the recommendation delivery mechanism. This intermediary layer handles the complex confidence calculations using pre-computed data, shielding the real-time recommendation system from complexity while maintaining speed

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20210374832A1Automatically determining in real-time a triggering model for personalized recommendations

Publication Date: 2021.12.02 WALMART APOLLO LLC

AI summary

A method including building a recommendation triggering model. The method can include receiving, via a user device of a user through a network, an add-to-cart command associated with an anchor item for the user. The method further can include determining, in real-time after receiving the add-to-cart command, a recommendation for one or more complementary items of the anchor item for the user. The method also can include determining, in real-time after determining the recommendation, a recommendation confidence for the recommendation. The method additionally can include after determining the recommendation confidence, when the recommendation confidence is positive, transmitting, in real-time through the network, the one or more complementary items to be presented to the user via the user device. The method likewise can include after determining the recommendation confidence, when the recommendation confidence is not positive, refraining from transmitting the one or more complementary items to the user. Other embodiments are disclosed.