Data mining is a career of extracting data, building models to analyze the data, and communicating with the demand department after the results are obtained.

An example: a bank's business unit has many potential loan applicants, the business unit to the data mining staff to put forward the demand, hope to be able to analyze which applicants are high-quality lending objects?

Data miners should first fully understand the needs of the division, and secondly, they should extract the relevant data from the database, and the work of extracting the data is sometimes done by the DBA. Well, now that you've got the historical data, your task is to build a model through the historical data to analyze what characteristics of applicants are capable of repaying the loan and not defaulting on the loan, and then use the built-in model to predict the new batch of applicants that we just got.

To be more specific: for example, if we find from historical data that applicants who are older than 35, male, married, with a family size greater than 3, and with incomes of $12,000 or more are ideal candidates for lending, then we use this criterion to qualify new applicants.

Of course the example I gave, for the sake of simplicity, is a very simple schematic example, the actual situation is much more complex, and will involve an individual's lending history, credit assessment, natural attributes, social attributes, asset assessment, etc. - that is to say, the data miners are trying to go through the huge amount of data in the database, sort out what is useful data, and then use these useful data to analyze the problems of other departments, help them solve the problem, or provide data basis for the development of the company

The upward direction of data mining is: data mining - product layer - Decision-making

java is part of the development, such as the development of software, interfaces, applications, etc., if a company needs to develop data mining software, then you need to data mining knowledge + java development capabilities, only in this case, you need to have both

But the general development of the data mining software companies are very few, the first need to consume a lot of manpower and resources, the second market has a lot of ready-made software. The second market has a lot of ready-made software, there is no need to develop.

If you want to engage in data mining, you must have:

Mathematical knowledge of data mining models, algorithms, and some data analysis software (SPSS, SAS, matlab, clementine)

Some database-related knowledge (oracle, mySQL)

Understanding of the market, the other departments demand

Of course, these are accumulated bit by bit, there is no need to overnight, especially the market, industry knowledge and understanding of the needs of other departments of the company is very important, which determines whether you can rise from the basic analysts to the product layer, decision-making layer, are to be accumulated in the actual work

As for the abandonment of java and so on, I think it is really not give up, because you have the foundation of java, must be able to come in handy, such as technical product managers (face book's zuckerberg and tencent's ma huateng are technical product managers), this kind of product manager is able to clearly grasp the development process of the product, as well as market knowledge. To summarize, nothing will be wasted, you learn all the things will be in the work of the field, just you encountered the situation is not enough more than enough complexity only