Visiting Frequency Estimation via Minhash Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining visiting frequencies among devices at a site are cumbersome and costly, requiring large storage for raw data and extensive processing time as the number of consumers increases.
Innovation Solution
The use of modified minhash functions to estimate visiting frequencies by generating a list of devices detected over a time period, selecting the smallest hash values, and storing only these minhashes, allowing for the determination of visiting frequencies and their distribution without storing raw data for extended periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raw data of all detected devices is stored up to a certain point in time to determine visiting frequency, then measurement precision of visiting frequency is improved, but storage cost and processing time increase significantly
Solution Approach 1:
The patent extracts only the essential information needed for visiting frequency determination by using hash functions to transform device identifiers into fixed-size hash values. Instead of storing complete raw device data, the system stores only hash values in a hash table, extracting the minimal necessary information while discarding redundant data, thereby reducing storage requirements while maintaining measurement capability
Solution Approach 2:
The patent transforms the device identifier parameter into a hash value parameter through hash function transformation. This parameter change converts variable-length device identifiers into fixed-length hash values, enabling efficient storage and comparison. The hash value serves as a compressed representation that preserves the ability to identify and count device visits without requiring storage of the original large-size device data
2Measurement precision
If raw data of all detected devices is stored to determine visiting frequency distribution, then measurement precision of visiting frequency distribution is improved, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing hash values of device identifiers as they are detected, rather than waiting until the end to process raw data. The hash table is continuously updated with device hash values and their visit counts during the monitoring period, so that when visiting frequency distribution is needed, the computation is already substantially complete, reducing final processing time
Solution Approach 2:
The patent replaces the mechanical process of parsing and analyzing complete raw device data with a computational hash-based system. Instead of mechanically searching through and comparing raw device identifiers, the system uses hash functions to rapidly compute and compare fixed-size values, substituting complex data processing mechanics with efficient computational operations that reduce processing time
3Loss of information
If more consumers/visitors are detected at the site, then the value of visiting frequency information increases, but storage cost and processing complexity increase
Solution Approach 1:
The patent creates a simplified copy of the device identification system by replacing complex device identifiers with hash value copies. The hash table stores copies of device information in compressed hash form rather than complete device data, enabling the system to handle larger numbers of consumers while maintaining manageable data structures and reduced complexity in data management operations
Data Source
AI summary
Disclosed are systems, methods, and computer-readable media for determining a distribution of various visiting frequencies among devices/elements detected at a site. A method includes generating a list representing a portion of devices detected at a site over a first time period based on wireless signals transmitted by the devices and received at an access point and determining a visiting frequency for each device represented in the list based on data representing devices previously detected at the site. The method further includes, based on the visiting frequency determined for each device represented in the list, estimating a distribution of each visiting frequency among the devices detected at the site over the first time period to yield a plurality of estimated distributions and storing the plurality of estimated distributions for further processing.