engleski

Automated Coding of Census Data

In an attempt to diminish the costs, increase timeliness, and improve the quality of the 1991 census data processing in Croatia and in Bosnia-Herzegovina, intelligent optical readers were purchased. Besides the fields containing coded numerical values, fourteen free-text variables were also included in the census. This article deals with the automated coding of these optically read free-texts. The Croatian language is highly inflected and nouns and adjectives change significantly with changes in case. This characteristic of the Croatian language has made the automated coding of census data more difficult. This paper discusses the coding algorithm and the results.

character recognition; coding algorithm; intelligent optic readers

nije evidentirano