Using Unsupervised Machine Learning for Data Quality. Application to Financial Governmental Data Integration

Data quality, means, that data are correct, reliable, accurate and valid to be used and to serve its purpose in a given context. Data quality is crucial to make right decisions and reports in every organization. However, huge volume of data produced by organizations or redundant and heterogeneous data integration make manual methods of data quality control difficult, for that using intelligent technologies like Machine Learning is essential to ensure data quality across the organization. In this paper, we present an unsupervised learning approach that aims to match similar names and group them in same cluster to correct data therefore ensure data quality. Our approach is validated in the context of financial data quality of taxpayers using scikit learn the machine learning library for the Python programming language.