| 摘要 |
The present invention relates to a method and a device for minimizing a database size of an n-gram language model. The method for minimizing the database size of the n-gram language model according to the present invention includes the steps of: generating one or more text data by analyzing an input signal; extracting a candidate text data in which a word corresponding to a dictionary database exists, from the one or more text data; calculating a hash value with respect to the candidate text data by using a hash function; and obtaining appearance frequency data corresponding to the hash value from an n-gram table corresponding to a word count of the candidate text data. The present invention has an effect that an index value is removed from the n-gram table by obtaining the appearance frequency data of the n-gram table through the hash value of the text data, thus a size of the n-gram table is reduced. |
