This is part of the CRDT Data Infrastructure project
Objective(s) and Scope
This project aims to develop improved data integration mechanisms for CRDT’s data assets, ensuring accurate and consistent data integration. This involves the development of an entity resolution framework and the implementation of probabilistic record linkage techniques to enhance the CRDT’s data integration processes.
The project will start with a comprehensive review of CRDT’s current data integration mechanisms and identification of areas for improvement. Based on this review, an entity resolution framework will be developed to ensure accurate identification of entities across different data sources. Probabilistic record linkage techniques will be implemented to link records from different data sources and provide a unified view of the data.
Industry best practices for entity resolution include:
Data Standardization: Normalizing and standardizing data from different sources before entity resolution to ensure consistency.
Data Quality Assessment: Assessing the quality of data from different sources to identify potential issues and resolve them before entity resolution.
Rule-Based Matching: Defining rules and criteria for entity resolution, such as exact matching or fuzzy matching, to improve accuracy and consistency.
Probabilistic Matching: Implementing probabilistic record linkage techniques, such as Jaro-Winkler or Levenshtein distance, Fellegi-Sunter model etc. to handle variations in data and improve match accuracy. Methods that leverage machine learning will also be implemented to improve matching performance.
Continuous Monitoring: Regularly monitoring and updating entity resolution processes and algorithms to ensure their ongoing effectiveness.
Integration with Data Management Processes: Integrating entity resolution into the CRDT’s overall data management processes to ensure consistency and accuracy of data integration.
- Start Date: 08/02/2020
- Status: Active