AI service "xenoBrain" that predicts market changes from economic news
We have developed and deployed "xenoBrain", a service that predicts future corporate performance and market trends by collecting economic news from all over the world and automatically analyzing the causal relationship of each information with AI. doing. As a function, it displays how certain economic news affects related industry companies, and extracts and displays news that may affect corporate performance such as own company, industry demand, material price, etc. It's something to do. We officially released this service in June 2019 and have already introduced it to major manufacturers and financial institutions to improve accuracy and service.
A conceptual diagram of service provision through a business alliance with Dow Jones' news content analysis platform "Dow Jones DNA".
The background to the establishment is that I worked for an audit corporation as an accountant while attending the Faculty of Commerce of Keio University, and after that I was involved in the development of economic information services. When he was in charge of business development of Asia's largest economic information platform "SPEEDA", which collects world economic information data on a user base and allows users to search, he was also involved in user interface and service design, while also AI. I was exchanging data analysis with engineers in the field, such as "what and how much can I do?" Isn't it possible to create epoch-making services by multiplying the financial knowledge that I have with technologies such as economic information and AI? Was born.
This is because, in a world where technology has developed so much, economic forecasting still relies heavily on the capabilities of specialists. It takes a considerable amount of time because it is necessary to read and analyze a huge amount of information such as financial data and statistical data. Moreover, quantitative data such as statistics, stock prices, and finances, which are expressed numerically, are the results of economic events and are merely past events, so the forecasting part must rely on the rules of thumb of experts. Was the actual situation.
If AI takes charge of such work, a very wide range of data will be available, from the climate and business sentiment of the raw material producing countries of the target industry, the international economic and political situation, to the performance of the company and products. Can be analyzed in an instant. Of course, highly skilled analysts may be able to perform deeper analysis based on experience, but AI should also be able to perform faster and deeper analysis by repeating learning. With that in mind, we established xenodata lab. In February 2016 and have been working on development.
Realize innovative services through economic analysis using natural language processing
Our technical characteristics are centered on natural language processing. "Natural language" is the words and sentences that we humans use for "machine language" that has structured rules like programs. These words are called unstructured data. For example, even if you try to make a machine read the sentence "The price of natural rubber has fallen, the price of raw materials has fallen, and the business performance has improved." It is only recognized as a list of characters such as "go", "mu", and so on. Therefore, first, part of speech is decomposed into nouns, particles, auxiliary verbs, etc. so that the composition of the sentence is understood. Furthermore, there is a raw material called "natural rubber", and it is necessary to mean that the "price" is "declining".
In addition to accumulating a huge amount of these patterns, we will improve the accuracy of analysis by using our unique technology to improve the efficiency of those patterns. By doing so, you will be able to accurately understand the content of each article and analyze interrelated phenomena. This is the reason why we are working on "economic analysis in natural language", which has not been realized yet, while the analysis of conventional quantitative data has been exhausted.
We will connect the links of countless economic events and predict future trends.
The actual service flow is to read economic news in real time and automatically generate a causal tree of companies related to that topic. Analysts may scrutinize the content and incorporate further insights. The currently recommended usage is to use this tree to discover causal relationships and variables that were not noticed by previous rules of thumb. If you are a manufacturer's business planner, you can discover new opportunities and risks from the relationships with other industries that you thought were not related to your company's business performance. If it is a financial institution or consulting firm, you can use it for risk investigation and management support.
It can also be used as a sales targeting list. For example, in the case of a real estate company, it can be used to comprehensively analyze performance, growth in the number of employees, trends in the industry, etc., and list companies that are likely to move or expand their offices.
Until now, "what will happen to the future" has been a major risk factor for business decision-making and management decisions. However, if the risk can be reduced by improving the prediction accuracy and automating, such as "xenoBrain", the quality of decision-making will improve, and not only the stabilization of management on a company-by-company basis, but also the industry as a whole, and eventually the world, will move forward. It may be an opportunity to move forward.
An example of a causal tree automatically generated by the news analysis function of "xenoBrain".
From prediction to search. "Democratization" of economic analysis and the future beyond
For us, 2019 was a milestone year when we released the official version of "xeno Brain". From here on, it's time to upgrade the version to improve the degree of perfection and scale it. KII has been supporting me since December last year, and it is very encouraging to be able to provide a wide range of consultations as I myself am required to make various decisions as a manager.
As a short- to medium-term goal in the future, I would like to change the way the act of "prediction" itself should be. For example, in the case of the manufacturing industry, instead of collecting various data such as raw materials and product markets and inputting them into Excel for calculation, just enter a search word such as "next year's steel price". I want to display the prediction result in. With "xenoBrain", the goal is to establish the world view that "the future is not something to analyze, but something to search for."
In addition, the economic analysis service using AI is called "Predictive Analytics" in the United States, and it is showing excitement such as the birth of a unicorn company. Japan is behind the scenes, but we dare to name this technology "Forecast Tech" and develop it as a new technological field. The strength of AI is that it can identify, generate, and predict while handling vast amounts of data. Prediction technology is the most difficult of these, but we would like to bring innovation to this area and create a major trend next to FinTech.
Furthermore, by being able to search the future of the economy, knowledge that was previously monopolized by experts as tacit knowledge will be visualized, and anyone will be able to utilize that knowledge. We call this "democratization of analysis." In addition to this great potential, my personal motivation is "I want to see with my own eyes a world where the future can be predicted with a single search." It must have a great impact on society as a whole. I am excited about such an image, and I would like to greatly expand this possibility.
[Link to official website]