The recognition of Laos organization name based on a cascaded conditional random fields

摘要

The recognition of Laos organization name is a difficult problem in the entity recognition of Laos language. This paper presents a algorithm of Laos organization name recognition model based on cascaded conditional random fields. The algorithm solved the recognition of easy entity such as person name and location name in the lower model of conditional random fields(CRFs) and served the recognition of complicated organization names on the higher CRFs. This paper designed a efficient feature template and automatic feature selection algorithm for the conditional random fields model of organization names. In the open test of a mass linguistic data, the recall rate reached 79.67%, precision rate reached 77.72%, F - measure reached 78.68%.