pminer_intro

PatentMiner (pminer.org) is a novel topic-driven patent analysis and mining system. It is designed for an in-depth analysis of patent activity at the topic-level. The main unique characteristics of the PatentMiner system that distinguish it from traditional patent search systems are as follows: (1) topic-driven modeling; (2) heterogeneous network co-ranking; (3) intelligent competitive analysis; and (4) patent summarization.

• Topic-driven modeling. The fundamental problem in most existing systems is that all patents are simply modeled based on keywords. In this work, we present a probabilistic model to simultaneously model the topical aspects of different objects in the heterogeneous patent network.

• Heterogeneous network co-ranking. When a company plans to enter a new market or brainstorm novel ideas, several typical questions are: what are the most active companies in this area? what are the most relevant patents? and who are the most prolific inventors? We propose a heterogeneous co-ranking algorithm to address these questions.

• Competitive analysis. It would be very helpful for a company to make right business strategies by identifying who are its competitors and what is the trend of a competitor’s technology development. We define four measures to identify competitors, topic-level competitors, and competitors’ evolutionary pattern based on the topic modeling results.

• Patent summarization. It is always expensive to digest the large number of patents returned by a search engine. We present a maximum coverage method to automatically generate a summary for the results of patent search.

Thanks to all members involved in this project!

The team members include:

Jie Tang, Principle Investigator and project leader, Associate Prof @ Department of Computer Science and Technology, Tsinghua University. His research interests include social network theories, data mining methodologies, machine learning algorithms, and semantic web technologies.

Yang Yang, Team leader, Ph.D. student @ DCST, Tsinghua University.

Yanting Zhao, Bachelor (graduated). Yanting was working on competitor analysis. She is now a master student in Columbia University.

Menglin Cai, Bachelor (graduated). Menglin was working on entity search. She is working in Shanghai.

Chengtao Li, Bachelor. Chengtao was working on interactive company search. He will graduate in the summer of 2014.

Xun Zheng, Bachelor. Xun was working on topic extraction. He is studying in CMU.

Tian Li, Bachelor (graduated). Tian was working on patent crawling & extraction. She is a master student in Dartmouth College.

Main contact: Jie Tang, jietang@tsinghua.edu.cn

Europe contact: Ghalem Ouadjed, Founder & CEO, SAS Eoweo, gouadjed@eoweo.com