Machine learning 00: How to get started with Python machine learning

We all know that machine learning is a very comprehensive research subject, which requires a high level of mathematics knowledge. Therefore, for non-academic professional programmers, if you want to get started machine learning, the best direction is to trigger from the practice.

PythonThe ecology I learned is very helpful for getting started with machine learning. So I hope to use this as a breakthrough machine learning.

I will record a series of learning and Practice records. Record the content of the main reference Youtube in the sentdex release of the video, interested readers can themselves fq to the tubing to see.

Here's how I'm going to Python learn from the Getting started machine.

Learn basic Python syntax

First I Python found the introductory tutorial on the official website, quickly over Python the basic grammar. I believe this is not a problem for someone with a little bit of programming foundation.

As a practice, I then Python implemented a command-line translation script. Get Python started with this.

Here's a lengthy process of setting up the Mac Python environment. In this article, I describe how to handle the system's own and installed Python versions.

Python machine learning related libraries

PythonThere are many libraries involved in machine learning, such as,,, and Theano TensorFlow PyTorch scikit-learn so on. Considering that scikit-learn sklearn machine learning is highly encapsulated and abstracted (hereafter abbreviated), it allows beginners to jump out of a mathematical nightmare for machine learning practice, and I choose it as a springboard for getting started.

In addition, we need to learn Python the following libraries for data processing or scientific calculation.

• numpy: Provides a powerful library of n-dimensional arrays and related operations, with reference to NumPy Quick Start notes.
• pandas: Provide a library of similar relational or tagged data structures, refer to Pandas Quick Start notes.
• scipy: A library that integrates many mathematical functions, please refer to the official documentation yourself.
• matplotlib: Tools for drawing data into images, refer to matplotlib Quick Start notes.

Departure machine learning Adventure Journey

sklearnProvides a lot of machine learning algorithm implementation, in the learning process I can not do a full study and coverage. After many searches, I found the Youtube sentdex released video "machine Learning with Python". At this point, I will also follow sentdex the footsteps of a step by stage to learn.

Follow-up articles are mainly reference videos, and combine their own understanding to carry out the necessary extensions.

sklearnfor the first time, you can read the official Tutorials documents.

The section "An Introduction to machines learning with Scikit-learn" will give you an idea of sklearn what this library can do, the basic concepts of machine learning, sklearn environment building, basic functions, and so on.

The "A tutorial on statistical-learning for Scientific Data Processing" section allows you to learn about the basic concepts of supervised learning and unsupervised learning.

In-depth principle

sklearnThe ability to provide the implementation of machine learning algorithms in a black box is beneficial for beginners. But if just stay here is obviously not enough, if you do not have a certain basic knowledge and principle, we can not model and model the display problem. So sklearn after learning the algorithm, be sure to consult the relevant documents, understand the knowledge and principles behind the algorithm.

This process should be the hardest and hopefully we will not stay at this step.

This article comes from a sync blog

Machine learning 00: How to get started with Python machine learning