The present invention addresses the problem of improving the accuracy of named entity recognition with respect to an unknown word not listed in a dictionary. A character string included in text data is divided into a plurality of tokens. A matching process is executed between a token array indicating a predetermined number of successive tokens among a plurality of tokens and dictionary information including a plurality of named entities, to search for a similar named entity, among the plurality of named entities, of which similarity to the token array is greater than or equal to a threshold. Matching information indicating a result of the matching process between the token array and the similar named entity is converted into first vector data. Input data is generated by using a plurality of pieces of vector data converted from a plurality of tokens and the first vector data, and a named entity recognition model for detecting a named entity is generated through machine learning using the input data.
展开▼
