Big data has become the meat and potatoes over the years, and everyone is talking about it, but do they really understand it? I don't think so.?

Many people seem to think that big data is actually on a large scale of data, after all, the name is so named. But, is big data really just that?

If you want to talk about big data, you have to put forward IBM's 5V theory: Volume (large amount), Velocity (high speed), Variety (diversity), Value (value) and Veracity (authenticity).

1, Volume (large number)

This feature is also well known, now is the era of big data, the data generated every day is extremely horrible, before the MB, GB has been far from enough to describe the current amount of data, and even can only be used to describe the ZB this kind of very large data unit. And to deal with big data, accordingly, must also use distributed computing can be realized.

2, Velocity (high speed)

Massive data requires sufficient storage space, but the processing speed must also be very fast, or the user experience will be greatly impacted, it is difficult to imagine Baidu search in the user's search keywords, 1 minute before the results. If the big data processing speed is not fast, this thing will really be a reality, and even more than ever before.

3, Variety (diversity)

The so-called big data, is not our traditional structured data, it should be said that the explosive growth of big data, in fact, is derived from the non-traditional unstructured data, that is, audio, video, pictures, geographic location and so on. These data are distinguished from the traditional two-dimensional structure, and have higher requirements for data processing, which is also an urgent problem in the era of big data.

4, Value (value)

Massive data does not mean massive value, not so. On the contrary, the data value density in the era of big data has become lower, with a big wave of sand to describe it is not too much. Then how to carry out efficient value mining? This requires the use of current machine algorithms to solve the problem, such as feature extraction, clustering algorithms, classification, such as automatic face recognition, something very simple for people, but very complex for machines.

5, Veracity (authenticity)

The above four points, I personally believe that is not the most important, the most important should be the authenticity, that is, the quality of the data. The quality of the good and bad, directly guarantee the final big data output cutoff is real and reliable. Many people will think that big data will be real, not so, take the advertising field, cheating traffic phenomenon can be seen everywhere. Therefore, big data will definitely be real, not so.