Big data can also predict the flu?!

According to the NHK website, as long as you log on to this magical website with your Twitter account, you can find people who may have a cold in a short period of time through tens of thousands of Twitter, and analyze the past and today's colds. In addition, the program will combine temperature and humidity changes to predict the future prevalence of colds, these will be through a "cold Calendar" to express. Currently, such services are unfolding in Japan. It is understood that the server by a pharmaceutical products manufacturing SS pharmaceutical company (エスエス production 薬) operation, the company said, through the analysis of the server, we can know how many people around their own symptoms of colds, and do a good job in advance prevention.

In addition to the SS pharmaceutical company, Japan's Kyoto University, the Department of the Wilderness has also run a website called "Influenza June" (インフルくん), which mainly functions in the search for cold information, can predict the flu in that area more serious, infection situation. The site will automatically classify the topic on Twitter and select a message that actually infects the flu and then edit it manually, after which the server is positioned by GPS to mark the location of the infected person on the map of Japan. If there is an increased trend of influenza in an area, the area will be marked red, reminding local people to take precautions.

The National Institute of Infectious Diseases in Japan will count and release data on influenza patients in about 5,000 health clinics nationwide. By contrast, the actual statistics obtained by the Institute are basically the same as those predicted on the website, so why are the results of large data very accurate?

The first is because through the network information analysis technology has made progress, already can through various messages automatically search the relevant data, and automatic classification. As Google now uses the technology, it uses the server to analyze more than 10 words that are closely related to influenza.

Another is the feature that is unique to large data. When the flu is worst, thousands of Tweets are released every day, even if there are some errors, but data analysis can also be used to analyze the accuracy of the data.

In the past, public agencies have had to delay at least a week in releasing flu information, and the information in some remote areas is not exactly the same, and now, through the Internet, these deficiencies can be effectively remedied.

The University of Kyoto, which runs the "flu June" website, says that more and more people are using social media, accumulating a lot of information that can be used to analyze and use large data to accurately predict the prevalence of infectious diseases.