The main purpose of this project was to develop a matching algorithm in python to fuzzy classify people from a customer list as positive or negative based on a messy positive and negative database with a confidence score.
Two approaches were used:
- Deterministic data matches were used to match the clear matches against the positive and negative list
- Probablistic data matches were used to match the fuzzy matches against the positive and nagative list with a confidence score
The main package used in the algorithm was Record linkage