Guojie on large data heat: recent Strictly

Academician of Chinese Academy of Engineering, Guojie

"In fact, bioinformatics, brain science, space science and other basic research has long adopted the number of PB-scale http://www.aliyun.com/zixun/aggregation/14294.html" > large data, but did not trigger a big wave of data. Big data is now being taken seriously, mainly because of its enormous economic value. ”

Recently, in Beijing, "large data background of the computer and Economic Development Forum," said Guojie, academician of Chinese Academy of Engineering, China to carry out large data research, we must take the enterprise as the main body, fully mobilize the enthusiasm of enterprises and economists.

An interesting example of the application of large data in the economic field is the prediction of stock price fluctuations through social networks. The social media monitoring platform, DataSift, monitors the link between Facebook's emotional bias on Twitter and the share price swings in the social networking site Facebook's IPO day. Facebook's share price began to fall after 25 minutes after Facebook opened with a gradual shift in sentiment toward the negative. And when the Twitter sentiment turned positive, Facebook's share price began to rebound in 8 minutes. Eventually, when the stock market closed and Twitter sentiment turned negative, Facebook's share price began to fall after 10 minutes. The agency concludes that every shift in sentiment on Twitter will affect the volatility of Facebook's share price.

"Predicting stock prices is a recognized and difficult thing to do, but big data analysis also works." "Other things that you used to think you couldn't do, maybe big data analysis could come in handy," Guojie said. Therefore, our urgent task is to practice more, practice knowledge, practice theory. It is only when the data analysis in each field is done quite deeply that the common problems of data science can be raised. ”

Although most units are still in the "small data" processing stage, Guojie that companies and government departments do not have to care too much about what they are analyzing is not "big data", not to spend energy on the various terms and definitions of unnecessary debate. "As long as there is a certain amount of time on the vertical accumulation, in the horizontal has a richer record details, through a number of sources of the same object collection of various data organic integration, careful and careful analysis of the data, it may produce great value." ”

Guojie also reminded that we should have a sober understanding of the technical foundation behind large data in China. Since our country has the largest population and the largest number of netizens, many people estimate that our annual increment of new data should be ranked among the best in the world. But in terms of the amount of data, China is not a big data country. According to the McKinsey report, the number of new data in the United States in 2010 was 3500PB, Europe was 2000PB and China was only 250PB. In other words, China's new annual data volume is only 7% of the United States, Europe's 12%.

"Of course, the key to large data is not the size of the data, but the ability to obtain effective information and the level of data analysis." "But in the Big data field, China and the world advanced countries are not the same starting line," Guojie said. ”

There are a lot of meetings involving big data, but guojie that governments, businesses and academia may not really understand the significance of big data. "Many people often estimate the benefits of transformative technology as being overly optimistic and short term estimates." ”

His own judgment is that, like cloud computing and the Internet of things, the market for big data in recent years is not big, and the world is about tens of billions of dollars. However, the growth of the big data market will be several times that of the traditional IT industry. "For the recent big data market can not be overestimated, can not be fooled themselves." The development of large data research and application must have a long-term vision. ”

(Responsible editor: The good of the Legacy)