Unsupervised ML Consumer Behavior Vector Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing geolocation analytics systems fail to accurately segment audiences based on consumer behavior due to insufficient specificity, poor correlation between location and behavior, and the inability to process large-scale consumer data effectively, leading to disregarded meaningful information and inadequate audience classification.

Innovation Solution

The development of an unsupervised machine learning model that discovers psychographic segments of consumers by converting consumer-behavior records into vectors and classifying them into specific segments, allowing for the prediction of behavior associated with these segments, even in high-dimensional and diverse data sets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If geolocation analytics systems use traditional audience classification methods, then the system complexity remains low, but the audience segmentation accuracy is insufficient and cannot effectively capture consumer behavior patterns

Engineering Contradiction:

Improveaudience segmentation accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transforms consumer behavior data from traditional tabular formats into vector representations in a multi-dimensional space, fundamentally changing the data parameter structure. This enables the application of geometric and algebraic operations to identify psychographic segments, significantly improving segmentation accuracy while managing complexity through mathematical transformations rather than complex rule-based systems

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces traditional mechanical classification systems (rule-based segmentation) with unsupervised machine learning models that automatically discover patterns in consumer behavior vectors. This substitution eliminates the need for manual segmentation rules and achieves superior accuracy through algorithmic pattern recognition

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Loss of information

If the system processes large-scale consumer data sets with high-dimensional features, then the comprehensiveness of consumer behavior analysis improves, but the computational complexity and data processing requirements increase significantly

Engineering Contradiction:

Improvecomprehensiveness of consumer behavior dataVSAvoiddata processing complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent maps consumer behavior records into a multi-dimensional vector space where each dimension represents a different behavior attribute. This dimensional transformation allows the system to process and analyze high-dimensional data using geometric operations, maintaining comprehensiveness while enabling efficient computation through vector algebra rather than traditional data processing methods

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Reliability

If traditional audience classification methods are used, then the implementation is straightforward, but the correlation between location data and consumer behavior is poor and lacks specificity

Engineering Contradiction:

Improvecorrelation between location and behaviorVSAvoidimplementation ease

Core Design Contradiction:

ReliabilityVSEase of manufacture

Solution Approach 1:

The patent segments consumer behavior data into distinct psychographic segments through unsupervised learning, creating homogeneous groups with strong internal correlations. This segmentation transforms the weak overall correlation into strong segment-specific correlations, allowing location data to be meaningfully interpreted within each segment context

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces consumer behavior vectors as an intermediary representation between raw location data and audience classification. These vectors serve as a mediator that captures the relationship between location and behavior patterns, enabling more reliable and specific audience segmentation than direct classification methods

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11238473B2Inferring consumer affinities based on shopping behaviors with unsupervised machine learning models

Publication Date: 2022.02.01 PLACEIQ

AI summary

Provided is a process of discovering psychographic segments of consumers with unsupervised machine learning, the process including: obtaining a first set of consumer-behavior records; converting the first set of consumer-behavior records into respective consumer-behavior vectors; determining psychographic segments of consumers by training an unsupervised machine learning model with the first set of consumer-behavior vectors; obtaining a second set of consumer-behavior records after determining the psychographic segments of consumers; converting the second set of consumer-behavior records into respective consumer-behavior vectors; classifying the second set of consumer-behavior vectors as each belonging to at least a respective one of psychographic segments with the trained machine learning model; and predicting based on the classification a likelihood of the respective consumer engaging in behavior associated with a corresponding one of the psychographic segments.