베이지안 텍스트 마이닝과 시각화를 이용한 특허 빅데이터 분석

The volume and variety are typical characteristics of big data. Compared to the large amount of data stored in relational databases, big data is stored by various data types such as texts, numbers, and pictures with unstructured data form. Patent document data also has a very large size and a patent document consists of various types of data such as the inventor's name, application date, name of the invention, technical summary, claims, and drawings. Due to the nature of the patent system which grants the inventors exclusive rights to patent technology registered for a certain period of time, much information about the developed technology is contained in the patent document. Therefore, patent big data analysis is an essential process for understanding technology. This study proposes an integrated analysis method for analyzing patent big data. The proposed method combines text mining and Bayesian inference to study the whole process from preprocessing to analysis and utilization of patent big data. We analyze the patent big data by newly combining Bayesian visualization and Bayesian regression. To show the performance evaluation of the proposed method, we collects patent documents related to artificial intelligence technology and uses them for analysis