Learning structural changes from text data





Who is this presentation for?
- Those interested in applying natural language processing techniques to business or policy analysis
Level
IntermediateDescription
There are volumes of text data in the digital era. However, this data is unstructured and sequential, which makes it difficult to extract meaningful information from it. Weifeng Zhong presents a novel method that uses deep learning to read large volumes of text and detect subtle, structural changes embedded in it.
As a first use case, economists Julian Chan and Weifeng Zhong developed the Policy Change Index (PCI) for China, an algorithm that can predict China’s potential policy changes. Using the information in two million articles in the government’s official newspaper, the People’s Daily, the PCI was able to predict many critical shifts in Beijing’s policy making. (More details about the PCI project can be found at Policychangeindex.org.)
The construction of the PCI does not require the understanding of the Chinese text—a “language-free” feature—which suggests a wide range of applications in other contexts, such as detecting changes in corporate announcements, legal documents, and political statements, among other arenas.
Prerequisite knowledge
- General knowledge about neural networks and natural language processing
What you'll learn
- Discover a new way to uncover hidden, structural changes in time series text data

Weifeng Zhong
Mercatus Center at George Mason University
Weifeng Zhong is a senior research fellow at the Mercatus Center at George Mason University. His work focuses on bridging the field of natural language processing and machine learning to economic policy studies. His other research interests include the political economy, US-China economic relations, and China’s economic issues. Weifeng is a core maintainer of the open source Policy Change Index (PCI) project, a framework that uses machine learning to “read” large volumes of text and detect subtle, structural changes embedded in it. As a first use case, the PCI for China is an algorithm that can predict China’s policy changes using the information in the government’s official newspaper. The PCI framework has received significant academic interest and media coverage. The resources of this project are freely available at Policychangeindex.org. Weifeng has been published in a variety of scholarly journals, including the Journal of Institutional and Theoretical Economics. His research and writings have been featured in the Financial Times, Foreign Affairs, The National Interest, Real Clear Markets, Real Clear Politics, the South China Morning Post, and the Wall Street Journal, among others.
Presented by
Elite Sponsors
Strategic Sponsor
Exabyte Sponsor
Impact Sponsor
Contact us
confreg@oreilly.com
For conference registration information and customer service
partners@oreilly.com
For more information on community discounts and trade opportunities with O’Reilly conferences
aisponsorships@oreilly.com
For information on exhibiting or sponsoring a conference
pr@oreilly.com
For media/analyst press inquires