.png)
Discover the capabilities of DarkMatch. Contact us for more information.
John Smith Jr.
123 Main St
01/01/1980
Applied for a mortgage in 2023
John Smith
123 Main St
01/01/1955
Each data point (name, address, birthdate, loan application date) would be transformed into a vector representation. This process captures the semantic meaning of the data, allowing the model to understand relationships beyond exact matches.
The transformer's attention mechanism would analyze the entire sequence of vectors for each record. It would recognize the significance of the age difference (25 years) between the two records.
By analyzing the loan application date (2023) in Record A, the model could infer that a person born in 1980 is more likely to be applying for a mortgage than someone born in 1955. This is based on general knowledge about life stages and financial behavior.
The combination of the age difference, the shared address, and the loan application context would lead the model to assign a higher probability to Record A being "John Smith Jr." and Record B being his father, "John Smith Sr."
Mary Jones
123 Oak Ave
Anytown, USA
Purchased Irish folk music CDs, Guinness beer, and Aran sweaters
Joined the "Celtic Heritage Club" loyalty program
Mary Smith
456 Elm St
Anytown, USA
Purchased a subscription to "Irish History Magazine"
Booked a guided tour of Dublin through the company's travel agency
Each data point (name, address, purchases, loyalty program) is transformed into a vector representation. This process captures the semantic meaning of the data, allowing the model to understand relationships and patterns.
The transformer's attention mechanism would analyze the relationships between the vector representations of various data points. It would detect a strong connection between the purchase history and loyalty program in Record A, all pointing to an interest in Irish culture. Similarly, the magazine subscription and the Dublin tour in Record B would be linked to the same cultural theme.
The model would recognize that both individuals, despite the name difference, exhibit a strong affinity for Irish culture. This shared interest, inferred from the context of their activities, would significantly increase the likelihood of them being the same person.
While not definitive, the consistent pattern of Irish-related purchases and activities could suggest that both Mary Jones and Mary Smith have Irish heritage or a strong connection to Irish culture. This inferred nationality could be added as a contextual attribute to their unified profile.
Contact our team today