Study hard and surpass yourself.
1. Build the environment
In response to the needs of my work, I started a long journey of machine learning, and I began to look confused. No one in the company knew this, so I could only bite the bullet and develop my own java. The leader suggested using python, so a python runtime environment was built based on sublime text3, and version 2.7.14 was selected.
Installation steps: https://my.oschina.net/wangzonghui/blog/1603104
2. Algorithm research
The company itself makes big data, extracts part of the data from the platform, and starts algorithm research. The leader stipulates that decision tree analysis is used first.
After a search on the Internet, there are math library and scikit-learn library, and the implementation methods of the two libraries are studied respectively. In the end, the rigid implementation of the decision tree is made. I won't go into details, there are personal blogs.
3. Conclusion
As far as the development of artificial intelligence in python is concerned, scikit-learn is better, but there are many algorithms, all of which have been optimized.
4. Reflection
A programmer who does not have a high degree of education and a high number of scumbags learns machine learning. He should focus on the code first, and increase the programming ability of the library or software according to the specific example. work, then implement the optimization, and then consider the adjustment of the necessary parameters. Build confidence step by step to learn better.