Data Mining Theory and Practice-how to link theory with practice and how to use Theory

Preface:

I bought the warrants (580013) for all the money in the morning (and I am still a student, and I don't have much money ...)... in just 20 minutes-10% .. heartbleed... I think about what I have done recently. It is estimated that most of my time has been spent on stocks, and I am almost a month away from my business .. what is business? I don't know. At least this blog is an algorithm ....

Bytes ----------------------------------------------------------------------------------------------------

The connection between theory and practice... my mathematical algorithms and the stock market estimated that there will not be anything substantive in the near future... let's talk about the combination of data mining and AI theory and practice ..

In my opinion, data mining = statistical methods + computers (databases) + business knowledge ....

Now Bi seems to be very popular, but in my opinion, it's just that the report system adds a simple statistical analysis .. however, this progress is good... but there are also some exaggerated places .. that's the covariance in the statistics .. what is the computational complexity of correlation coefficient calculation? You calculate the covariance of the 10000x10000 data. What is the computational complexity of the correlation coefficient? Try .. but what is telecom data mining? It means that data is *** million every day .. why don't you know .. if you can't even calculate the most basic statistics like this, it's intelligent .. (intelligence is high and low .. compared with the previous one, it is indeed smart ).. the method to reduce the data volume is sampling .. but how to smoke is another problem .. it is also a problem to check whether it is correct. in practice, the problem may be a set .. or a set of multiple .. it seems like a system...

Theory and Practice: I have done theory before and have done practical things. Of course, the conversion in this process is very difficult...

The theory author said to a number: To do a regression first and to analyze the error. in their eyes, these are very easy. the MATLAB command is OK, but if the data reaches 100,000, the MATLAB will die. The examples in their academic papers are only a dozen or even a few hundred. If there is more data tables, dozens of businesses are required, the layout on each page is also cheap .. so they naturally think that .. computing is not a problem. I used to do this too .. later I figured out my Sb .. A month later ..

Coder: The program theory and logic of the algorithm are complex .. they thought that this was not what we did. It's better to build a website with a variety of mature tools to make money .. why bother to do this ..

Business personnel: They think that computers are the most amazing. If they want to get what they want, they will come up with a *** requirement that they can be used immediately .. don't they think this can be achieved?

The importance of the interface is also suddenly determined ..

Interface can be theoretically connected to the actual and business... it seems very difficult to do so.

Rare talents have value...