Fujitsu Develops Technology Using Open Data and AI to Support Decision Making by Regulatory Authorities

What Is Needed to Use Valuable Data in Business?

In recent years, the spread of open data has increased the amount of data available to companies. For this reason, many companies face the issues of how to use large amounts of data in their businesses, as well as how to extract valuable information from the huge quantities of data.

Fujitsu Laboratories of Europe is engaged in R&D of technologies necessary to make open data and big data useful in business.

How Open Data and AI Can Change Regulatory Operations

One technology under development is Advanced Data Analytics for regulatory authorities. This technology aims to obtain data from different datasets, including private information held by regulatory authorities, open data (e.g., financial statements), and social media data, as well as to provide valuable data through advanced data analysis using machine learning, graph networks, AI engines, and other analysis tools.

"Advanced Data Analytics is based on the idea of creating a knowledge base within the system. It implements machine learning on top of the generated knowledge base and proposes relationships between data sets to generate new data," said Masatomo Goto, Manager of the Data Analytics Research Division at Fujitsu Laboratories of Europe.

By adding open data to the data already held by regulatory authorities, one can find hidden human relationships, relationships between companies, and relationships between specific regions and companies that had not been identified before, all of which are useful in regulatory operations. Moreover, it also becomes possible to analyze and detect abnormal market movements from huge quantities of historical stock purchase data.

The key members of a company can be displayed based on financial reports and corporate information.

Stock prices, company news, tweets, and other information can be viewed in one application.

Significantly Reducing Data Cleansing Time by Utilizing a Knowledge Base

After performing an analysis across data sets, the acquired data cannot be used as is. Up to the present, data analysts have spent more than 80% of their time performing data cleansing, leaving them insufficient time to conduct data analysis.
For example, our company can be described in different ways depending on the data set, such as "Fujitsu," "Fujitsu Ltd.," or "Fujitsu Limited." Computers usually process this information as if the names were referred to different companies. Advanced Data Analytics enables data cleansing through AI-based analysis to determine whether companies are in fact the same organization by referencing external information, such as a company’s address and president’s name.

"Currently, we are proceeding with development for regulatory authorities, but in the future, we would like to make this technology available to customers in other industries who need to analyze huge volumes of data, such as banks and regular companies. In addition, we have started to investigate whether we can identify cross-border relationships by integrating open data sets around the world. At present, we are working on experimental integration of data sets between Japan and Spain." (Goto)

Developing Technology for Every Case that Requires Analysis of Huge Amounts of Data

"We believe this technology can be used to support the decision-making of corporate managers and credit investigators, as well as to detect relationships between organized crime and companies. Therefore, we would like to offer this service in a more customer-friendly way through co-creation with customers." (Goto)

In the future, Advanced Data Analytics is planned to be released as an API. Fujitsu Laboratories of Europe is engaged in R&D with the aim of developing technologies that meet not only the needs of regulatory operations, but also other types of customers who need to analyze huge quantities of data.

Masatomo Goto
Manager, Data Analytics Research Division,
Fujitsu Laboratories of Europe
After joining the company, Goto engaged in R&D mainly to standardize data description languages such as SGML, XML, and XBRL. Since 2014, he has served as XBRL International Standardization Officer. He worked in Silicon Valley and New York from 2001 to 2009, and he has been working in London since 2012.