Croatian American Population Estimate: Based on Big Data

This is an estimate of Croatian American population based on “Big Data”. The total number of Croatian Americans was substantiated by comparing relatively large databases (big data). Standard Information Retrieval procedure was used, based on precision and recall when comparing the databases. Telephone directories in Croatia and in the US were used as large datasets to ensure the quality of the research. A research sample was a selected group of 183 typical Croatian family names that represents about 9% of the total population in Croatia. A large number of people was found in the US telephone directories (white pages) with the selected family names. The database of the Croatian Fraternal Union (CFU) reveals how many family names were changed due to marriages and other reasons (used as true positive data for recall index). A special model was created, that determines the loss of family names when having a “perfect elastic impact” with members of other communities. The result was calculated as a direct product of precision, recall and linear estimation. Error estimation was not completed, neither was the research repeated on different datasets. The book was the Author’s Final Thesis at the Croatian Diplomatic Academy in 2001.