| 摘要 |
The invention relates to a method for the computerized detection of a character string similarity between a first character string (1) comprising a plurality of characters (2) that is electronically acquired as a search term and at least one second character string (3) comprising a plurality of characters (2) that is saved in a database (6, 7, 8) as a proprietary name. The invention is characterized in that at least one representation of the character string (1) of the search term and at least one representation the character string (3) of the proprietary name is automatically split up into n-grams of identical character number, every n-gram having at least two characters (2). The n-grams of the search term and the n-grams of the proprietary name are automatically compared with each other to detect the character string similarity. |
