Trajectory Analysis for Shopping Group Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing shopper behavior in retail environments struggle to accurately identify and analyze shopping groups, as group behavior can be complex and contaminated by independent movements, leading to inaccurate data on product category associations and decision-making processes.
Innovation Solution
A method and system that uses probabilistic graphical models, specifically Hidden Markov Models, to analyze shopper trajectories and identify atomic behaviors such as staying together, walking together, splitting, and merging, to determine group membership and improve the accuracy of shopper behavior data by clustering trajectories based on dynamic interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If trajectories of shopping groups are analyzed independently without group membership information, then the analysis process is simple and fast, but the accuracy of product category association data is corrupted due to independent movements of group members
Solution Approach 1:
The patent segments the trajectory analysis into two distinct stages: first identifying shopping groups by detecting atomic behaviors (staying together, walking together, splitting, merging), then analyzing product category associations within each identified group. This segmentation allows the system to handle group dynamics complexity separately from the association analysis, improving accuracy without proportionally increasing overall system complexity.
Solution Approach 2:
The patent performs preliminary group identification and atomic behavior detection before conducting the product category association analysis. By pre-clustering trajectories into shopping groups based on their spatial-temporal patterns, the system prepares the data structure in advance, enabling more accurate association analysis while managing computational complexity through staged processing.
2Measurement precision
If trajectories of shopping groups are combined together to improve data accuracy, then the accuracy of shopper behavior data increases, but the complexity of detecting and measuring group behaviors increases
Solution Approach 1:
The patent breaks down complex group behaviors into four atomic behaviors (staying together, walking together, splitting, merging) that can be detected and measured independently. Each atomic behavior has specific detection criteria, making the measurement process more manageable and accurate while still capturing the full complexity of group dynamics through composition of these atomic units.
Solution Approach 2:
The patent uses a dynamic probabilistic graphical model that adapts to changing group behaviors over time. The model updates probabilities of atomic behaviors based on observed trajectory patterns, allowing the system to detect and measure group behaviors dynamically rather than using fixed rules, thereby improving accuracy while managing detection complexity through adaptive algorithms.
3Loss of information
If group behavior data is collected and analyzed, then richer marketing data on decision-making processes is obtained, but the complexity of the data processing system increases
Solution Approach 1:
The patent segments the data processing into distinct modules: group identification module (detecting atomic behaviors), trajectory analysis module (analyzing spatial patterns), and marketing insights module (deriving decision-making patterns). This modular segmentation allows rich marketing data to be extracted by combining results from simpler, specialized modules rather than requiring a single complex processing system.
Solution Approach 2:
The patent introduces atomic behaviors as intermediary representations that bridge raw trajectory data and marketing insights. Atomic behaviors serve as intermediate features that capture group dynamics in a simplified form, enabling the derivation of rich marketing data about decision-making processes without requiring direct analysis of complex raw trajectory data, thus reducing overall system complexity.
Data Source
AI summary
The present invention is a method and system for detecting shopping groups based on the dynamic relation between shoppers' trajectories. First, shopper trajectories are generated using video images or positioning devices. Then, group behavior features are extracted from a candidate shopping group trajectory pair. The group behavior features of a given pair of trajectories are typically the changes in positional differences and the average speed of these trajectories. From a model of shopping group behavior, a given pair of candidate shopping group trajectories is analyzed to determine the group score—the likelihood of the pair of shoppers indeed belonging to the same group. Lastly, the system utilizes graph segmentation framework to find clusters where trajectories belonging to each cluster have tight group scores with each other.