The Research on "Big Data" is the exploration of "Relevance ".

Unlike the traditional logical reasoning research, big data research conducts statistical search, comparison, clustering, and classification analysis on a large amount of data, therefore, it inherits some characteristics of statistical science. Statistics focuses on the correlation or correlation of data. The so-called "correlation" refers to the regularity between values of two or more variables. The purpose of "correlation analysis" is to find the hidden relationship networks (associated networks) in the dataset. In general, parameters such as support, reliability, and interest degree are used to reflect the correlation. The two data items A and B are correlated. They only indicate that A and B have an influence on each other when they are de-valued. They cannot tell us that a must have B, or, in turn, B must have.

Strictly speaking, statistics cannot test the logical causal relationship. For example, according to the statistical results, it can be said that "the incidence of lung cancer in the smoking population is several times higher than that in the non-smoking population", but the statistical results cannot draw a logical conclusion that "Smoking is carcinogenic. Chen xiyong, one of the founders in China's probability statistics field, used this example to illustrate the characteristics of Statistics during his lifetime. He said: if such a gene leads to two things at the same time: smoking and lung cancer. This assumption can also explain the above statistical results. In this assumption, this gene and cancer are causal, while smoking and lung cancer are correlated. Statistical correlation may lead to the illusion that the results are the cause. For example, the statistical results show that before the rain, swallows often go low. From the time sequence, the relationship between the two may indicate that the Swallow Falls low, which is the cause of the rain. In fact, it is the reason that the Swallow Falls low.

I used to think about the world with logical thinking. The occurrence of things has a causal relationship. In fact, you can change the mode and pay more attention to "related ".