Abstract: Aiming at the problems of unclear entity boundaries, many specialised vocabularies and multiple meanings in patent text data, this paper proposes a named entity recognition algorithm based ...