This thesis focuses on verification of effectiveness of various
ways of completing texts with diacritics in Czech language. The
main part of this thesis aims to find words which are ambiguous
without diacritics and also contains a division into categories
and a created set of rules for automatic completion of
diacritics. This set is tested on a corpus and its success rate
is compared to other methods of automatic completion of
diacritics. The following analysis of results determines
advantages and disadvantages of individual approaches and
suggests a possible improvement.