System and method for patent information niose detection based on deep learning

摘要

Disclosed is a deep learning-based patent information noise detection method and system thereof. In the deep learning-based patent information noise detection method, the noise detection system builds a word-to-vector engine or a centen-to-vector engine through deep learning of patent documents from the patent information DB, and the noise detection system is the patent desired by the user. Receiving a search condition for searching a document, specifying a raw data patent list including patent list data corresponding to the search condition received by the noise detection system, and the noise detection system being included in a specific patent list Converting each of the patent documents into a word vector set or a sense vector set through a word to vector engine or a sense to vector engine, and the patent list based on the converted word vector set or sense vector set for each patent document It includes the step of detecting a noise document among the patent documents included in.