For beginners of python and machine learning, I want to know how to develop programs independently?

I am learning python. I used to study at a university. C. I can only take the test. I am engaged in the retail industry, but I really love computer and programming. I bought a python learning manual and a core python programming book. However, after reading the strings, dictionaries, and lists, I am confused. Although the above exercises can be done, I don't know how to use them to build a program. Many codes and function definitions are completely unknown, even if you understand the operating principles of algorithms, you cannot write your own code independently. It can only be written based on the code in the book. I want to know that I am learning python. I used to study at a university. C. I can only take the test. I am engaged in the retail industry, but I really love computer and programming. I bought a python learning manual and a core python programming book. However, after reading the strings, dictionaries, and lists, I am confused. Although the above exercises can be done, I don't know how to use them to build a program. Many codes and function definitions are completely unknown, even if you understand the operating principles of algorithms, you cannot write your own code independently. It can only be written based on the code in the book. I want to know how to turn this knowledge into the ability to write your own code. I want to work on machine learning or data mining in the future. Reply content: first, practice Python. After completing the tutorial, you can practice Python with some small demands in your life, find some code from others, and try to rewrite it again. For more information, see:
What are Python's one thousand-line classic trainer projects? -Programming
How do you learn Python by yourself? -Investigation issues
What are the recommended Python trainer projects? -Programming

When learning Python, you can look at machine learning or data mining books, such:
Which books are recommended for data analysis and mining? -Book recommendation
How to systematically learn Data Mining? -Data Mining

Then, I will try to implement some algorithms and participate in Alibaba's Tianchi competition or kaggle trainer. Even if I get started, I will try to find a relevant job, it is easy to get into the road. Some graduate students engaged in machine learning and scientific computing will encounter many difficulties when they directly go to the python third-party library to write code, it is recommended to supplement basic knowledge.
Whether you can write code to solve the problem is to determine whether you have learned programming.

We recommend that you not only learn the syntax, but also the Computational Thinking, programming ideas, and solutions to the problem.

Read the article written by Huang ge.
How to crack the layer of python Programming
Article/pythonstudy. md at master · pythonpeixun/article · GitHub

Three languages: python2, php, and go
Article/jdstb. md at master · pythonpeixun/article · GitHub
A small piece of code illustrates the use of @ property decorators
A small piece of code illustrates the use of @ property decorators


How to crack the second layer of python Programming
How to crack the second layer of python Programming

How to crack the third layer of python Programming
How to crack the third layer of python Programming



Huangge python Remote Video Training Course
Article/index. md at master · pythonpeixun/article · GitHub

Yellow brother python Training Workshop video playback address
Article/python_shiping.md at master · pythonpeixun/article · GitHub I recommend you a book "Collective smart programming".
All the examples in this section are written in python. You may learn a lot from them by reading all the code.

Compared with python, this book makes me feel more like the idea you need and uses programming to solve the problem.

Finally, when you grow up, you may be questioned. Ignore it and do what you want to do. Your success is the most powerful weapon to eliminate doubts.

Cheering is not a blow to your self-confidence
In China, almost all those who want to do machine learning require graduate students
And you have all worked, and you can only recruit companies. You may not be able to participate in the Alibaba big data competition.

During the interview, the interviewer asked you a quick sort, binary tree or something, and you probably won't be able to write it out.
The interviewer asked you again about the operating system, computer network, massive data, and so on. What do you think you should do?
Since it is machine learning, at least a few questions will be given during the interview to deduce the mathematical process of machine learning, and then I will talk about statistical probability theory or something. Then I will ask a few questions about C ++ or Java, then, let's talk about Hadoop, Spark, and Storm.

If you want to go to a small company to do data excavator learning when I did not say (but a small company to do machine learning is basically directly dug from a large company), if you want to go to a medium or large company, I think that unless you have made a good score on KDD or Kaggle, you are still ready to take the postgraduate entrance exam, or you can just think about it at ordinary times (after all, there are not many jobs in China related to data mining, and there is little job demand)

Finally, it is recommended that, if you really like Python and machine learning, you should use python to write the code in machine learning practice and understand some simple principles of clustering classification algorithms, you can write kmeans And Naive Bayes, because these libraries all have third-party libraries. If you do not need a large amount of data, you can directly use the sklearn library, which is especially convenient. If there is a large amount of data to be distributed, I only use mapreduce to write data that is not distributed and there are many ready-made libraries. Therefore, the machine learning algorithm mainly needs to understand the principle and know how to apply each algorithm.

There are so many classification algorithms. We need to understand the principles and connections of each algorithm when we use specific datasets. For example, LDA is not supported when we encounter data classification with non-normal distribution, decision Trees may be used for discrete data classification. These are not absolute and can be used easily. There are so many machine learning algorithms that it is too difficult to understand all of them. Each category has some basic baseline, which one should be used for specific research. For example, in the recommendation system, after basic computing, the competition still fails to achieve good results, because data preprocessing is very important, whether in the competition or in the project, data preprocessing requires many machine learning algorithms.

I don't know much about my work, but what I learned after three months of practice is that machine learning is not used much. I often find rules and filter data, which is endless... You are not connected to Python at the level. We recommend that you do not want to write programs or machine learning. First, you can understand the Py syntax. If you are familiar with the features of some OOP languages, you can proceed with your plan. You just read the list dictionary, and there is still a long way to go. You can take a look at the MOOC tutorial. Every knowledge point has exercises. It seems that there is no foundation and you don't know where to start.
You can use sklearn to apply Machine Learning Algorithms on small-scale data.
Or you can take a look at the practices in machine learning practice and collective programming wisdom. You can use checkio to answer questions. This is suitable for beginners and is moderately difficult. After a question is completed, looking at others' answers is quite rewarding.