Product functional information based automatic patent classification: Method and experimental studies. (July 2017)

• Record Type: Journal Article
• Authors: Li, Wen-qiang; Li, Yan; Chen, Jian; Hou, Chao-yi
• Is Part Of: Information systems. Volume 67(2017)
• Page Start: 71
• Page End: 82
• Publication Date: 2017-07
• Subjects: Innovation design -- Functional basis -- Patent text classification -- Naive Bayes -- EM algorithm; Database management -- Periodicals; Electronic data processing -- Periodicals; Bases de données -- Gestion -- Périodiques; Informatique -- Périodiques; Database management; Electronic data processing; Periodicals; 005.7
• DOI: 10.1016/j.is.2017.03.007
• Languages: English
• ISSNs: 0306-4379

Abstract

Highlights: This paper studied the automatic mining and classification of the functional information from the patent texts to support the product innovation design. The classification on the primary level functional groups has been carried out and the classifier with the best classification accuracy has been built through a couple of experiments. In overall, all three experiments obtained decent accurate classification results so it proved that the whole experiment design is an efficient attempt to extract the hidden patent information and classify them automatically. We can further study the experiment results and put forward some suggestions on the automatic patent text classification process. Abstract: In order to effectively extract the hidden information from the patent texts and to further provide this information to support the product innovation design process, this paper proposed an automatic patent classification method based on the functional basis and Naive Bayes theory. The functions of products are regarded as the innovation attributes, and the function co-reference relations of the patents in different areas are established. Patent classification methods are proposed based on the functions of products and the general steps of the patent classification process are proposed. In addition, three training methods are studied in the experiments, including multi-classification fully supervised training, multiple dichotomous supervised training and semi-supervised training. Through comparing and analyzing the experimental results, a patent text classifier is developed. In summary, this paper provides a general idea and the relevant technologies on how to build a patent knowledge space by automatically extracting and expanding the patent texts.