Big data is "beer + diapers"? Interpretation of the three-dimensional characteristics of "Da"

The concept of "big data" originally originated in the United States. Here the "big" is usually used to describe the three-dimensional characteristics of the data: the first dimension refers to the increasing volume of information, the second dimension refers to a wide variety of information, and the third dimension index becomes usable information and can be analyzed faster and faster.

As early as 1969, Wal-Mart, the world's retail giant, used computers to analyze consumer purchases, and found that male shoppers tended to "match" a few bottles of beer while buying diapers, and promoted the sale of beer and diapers. Today, this "beer + diaper" Data analysis has become a popular explanation for scientists "big data" technology classic case.

"Big Data" has a variety of definitions, the concept originally originated in the United States, by Cisco, Granville, Oracle, IBM and other companies to develop the initiative. The ' big ' here is usually used to describe the three-dimensional characteristics of the data: the first dimension refers to the increasing volume of information, now from TB to EB-level, and will jump to the zettabytes level; the second dimension refers to a wide variety of information; The third dimension refers to the fact that data becomes usable information and can be analyzed more and more quickly. In an interview with our correspondent, Dr. Yang Yi, a computing and communications expert at the University of California at Berkeley, stressed that "the big data that the media has been talking about recently, in fact, refers to the fourth dimension of large data, namely the value of data, which is mainly embodied in the intelligent analysis of data." ”

Yang Yi says intelligent analysis is an investment in future wisdom, and the ultimate goal of analysis is to make smarter decisions. Some people mistakenly think that large data is just an upgrade of data technology processing, that is, how to store and back up the owned data, but the real meaning of large data is to use new ways to analyze the data and make a reasonable explanation.

The information involved in large data is closely related to people. People-related information accounts for 90% of all the world's data, mainly including e-mail, video files, social networks, blog content, call center dialogue, etc., it is growing at an alarming rate, the annual composite growth rate of 62%. The application of large data will help decision-makers flexibly respond to the opportunities and challenges arising from the "Data Tsunami" in the real world.

"Human information" will cause the evolution of information Technology (IT). For years, the scientific community has been changing the technology of Information Technology (IT), the content of "T". For example, the introduction of mainframe computers, client servers, network supply (IP), cloud computing and other technologies. For the first time, large data technology has changed the information, that is, "I" in it, from the traditional information (information) "I" Toward the "I" of the Intelligent (FDI) development.

The "technical understanding" of human information requires the fundamental adoption of new methods and new technologies to provide insights, ideas and intuition for the information that human beings are constantly increasing. The future of information computing will be a large data model, resulting in the fundamental transformation of human social Information interaction, information applications based on intelligent analysis will be more large-scale interpretation of human beings, mining the value of human information.

When it comes to uncovering crime through information search, the offender can look for evidence from the suspect's e-mails; when trying to understand the customer base, marketers can search for and analyze all the information that the client exposes, either as a researcher's own database or as a message from an investigator exposed to a microblog or blog. This picture makes us see that information flows and analyses are becoming more complex in today's exploding societies.

The White House Science and Technology Policy Office released its big Data research and development plan last March 29, and set up a "Big data advanced Steering Group" to show that the United States has raised the opportunities and challenges of the big data technology revolution to a national strategic level.

Last August, a survey of 1105 U.S. companies and groups showed 63% per cent of respondents believed that only the implementation and use of large data technology to complete their own agency mission, 49% of respondents said they will increase their large data budget, 46% of the respondents plan to maintain at least their existing large data budget level.

Yang Yi told our correspondent that for developing countries like China, it should be easier to catch up with the international advanced level in the High-tech field, and should increase the intensity of the research on large data. Mainly includes two aspects, one is in the Big Data Technology Foundation cloud computing aspect, must exert the application development; second, in the data processing, not only must pay attention to the structural data (namely already digitized information), but also should pay attention to in the electronic mail, the blog and so on the non digitized humanities data analysis.

(Responsible editor: The good of the Legacy)