Mobile Recommendation System Using Sensor Context and Feedback Loops

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing profiling methods for mobile computing devices are not always accurate, as they rely on user-generated data that can be biased and non-evolutionary, failing to correct for psychological influences and providing non-personalized recommendations.

Innovation Solution

A system and method utilizing a mobile computing device with sensors to collect contextual data, combining it with personalized datasets, and using a self-learning algorithm to provide personalized recommendations, which checks user behavior and adjusts recommendations based on follow-up data to reduce bias and improve accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If user-generated content is used for profiling, then the system can obtain initial user preferences, but the profiling accuracy deteriorates due to biased and non-evolutionary data

Engineering Contradiction:
Improveuser preference informationVSAvoidprofiling accuracy
Core Design Contradiction:
Loss of informationVSMeasurement precision

Solution Approach 1:

The system implements feedback loops where recommendation outcomes are monitored and used to refine future recommendations. The central processing unit checks follow-up of transmitted recommendations and bases further recommendations on checked follow-up, creating a continuous improvement cycle that corrects biases in user-generated data over time

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system employs self-learning algorithms that automatically adjust and improve profiling accuracy without requiring manual intervention. The central processing unit autonomously combines personalized dataset information with contextual data, checks recommendation follow-up, and refines future recommendations based on observed user behavior patterns

Inventive Principle:
Principle #25Self-service

2Adaptability or versatility

If static personalized datasets are used, then the system can provide initial personalization, but the recommendations become non-evolutionary and fail to adapt to changing user preferences

Engineering Contradiction:
Improvepersonalization capabilityVSAvoidrecommendation evolution
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The system transforms static personalized datasets into dynamic, evolving profiles by continuously integrating contextual data from sensors and monitoring recommendation follow-up. The central processing unit dynamically adjusts recommendations based on real-time contextual information and observed user behavior, enabling the system to adapt to changing preferences while maintaining personalization

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system ensures continuous improvement of user profiling by maintaining ongoing data collection through sensors, continuous monitoring of recommendation outcomes, and persistent refinement of the personalized dataset. This continuous cycle keeps the personalization capability evolving and adaptive rather than static

Inventive Principle:
Principle #20Continuity of useful action

3Quantity of substance

If user-generated preference data is collected, then the system can build initial user profiles, but psychological biases in user input reduce recommendation reliability

Engineering Contradiction:
Improveuser preference dataVSAvoidrecommendation accuracy
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The system introduces contextual data from sensors as an intermediary layer between user-generated preferences and final recommendations. This intermediary contextual information helps validate and correct biased user input by providing objective environmental and behavioral data that the central processing unit combines with user preferences to produce more reliable recommendations

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2677484B1System and method for making personalised recommendations to a user of a mobile computing device, and computer program product
Publication Date: 2017.07.19 SENTIANCE
  • EP2677484B1 patent drawingFigure 1
  • EP2677484B1 patent drawingFigure 2
  • EP2677484B1 patent drawingFigure 3

AI summary

System for making personalised recommendations to a user of a mobile computing device. The system comprises at least one mobile computing device, a central processing unit, and a memory structure comprising a dataset of personalised data. The at least one mobile computing device is provided with a plurality of sensors arranged for collecting contextual data. The central processing unit comprises receiving means for receiving said collected contextual data. The central processing unit is arranged for combining information from the personalised dataset with the received contextual data, in order to obtain a personalised recommendation. The central processing unit is provided with transmitting means for transmitting the personalised recommendation to the user. The central processing unit is arranged for checking follow-up of the transmitted recommendation by the user by comparing the transmitted recommendation with contextual data received by the central processing unit. The central processing unit is further arranged for basing a further recommendation at least partly on said checked follow-up.