Please use this identifier to cite or link to this item:
Title: Understanding technology flux : analysing patent applications to understand economics activities
Authors: Nguyen, Dinh Phuc
Keywords: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Document and text processing
Issue Date: 2016
Abstract: Patents have become important intellectual properties of every company and may reflect market trend if analysed carefully. The number of patent applications is rapidly increasing worldwide, creating a great demand for a computer-assisted classification system to reduce human effort in categorizing new patent applications. In the first half of this project, a simple yet effective system was developed that helps categorize new patent applications using International Patent Classification (IPC) taxonomy. The system used Naïve Bayes (NB) classifier for its simplicity and effectiveness [1]. The NB classifier was trained and tested on a standard patents collection provided by World Intellectual Property Organization (WIPO) [2]. The highest accuracy achieved were 43.73% predicting the exact category in one guess and 65.19% predicting the exact category within 3 guesses. Patent analysis is an important step for every company that wants to understand market trend and economic landscape. In the second half of this project, a search engine was built upon Elasticsearch [3] with data crawled from Google patents database [4] and U.S Patent and Trademark Office (USPTO) [5] . The data can then be sought, visualized and analysed with Kibana [6]. In the end, approximately 2.7 million patents from 2005 till 2016 were crawled, indexed and made searchable via Kibana.
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
5.15 MBAdobe PDFView/Open

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.