1. The data mining is a subject or the clear thing you want to solve, including the association between the some factors, finding their relationship, and so on.
2. The machine learning is a technique to help us to solve the subject such as the data mining, robot, pattern recognition, and so on.
3. Data mining does not have to use machine learning, you can also use statistical methods.
4. In summary, machine learning is a technique in data mining.
According to my understanding, machine learning is closely connected with building artificial intelligence, where the machine itself will learn from the data to perform a desired function in future. It’s similar to human beings who learn to do things from experience. Whereas, conventional data mining is restricted to extracting information from the data by means of statistical techniques such a correlation, regression, factor analysis, etc. This requires explicit programs to be developed for each intended analysis.
Precisely, data mining helps the investigator (human beings) to understand the system or problem for proper decision making and execution. In machine learning the humans are substituted by machine (computer) which will learn by itself to take actions accordingly. Even though, this requires the development of algorithms for training the machine based on historical data, it nullifies the need of explicit programming (corresponding to each analysis) that is quite complex in some circumstances. Hence, machine learning becomes useful when you want to automate insights around very large datasets that would be too difficult for a human being to do on a recurring basis.
In order to train the machine we have to mine the data as deeper as.
So train the machine is machine learning and mining the data means data mining and I added one more term "deeper mining of the data" is one of the example of deep learning.
the machine learning is concentrating on algorithms as a mechanism and the data mining is concentrating on the data as a content, furthermore the algorithm needs the data to prove its validation and accuracy. the data mining needs algorithms to discover hidden or new patterns so all of them are overlapped.