1 Introduction 1.1 Wrong idea of machine learning
- Be sure to know a lot about Python programming and Python syntax
- Learn more about the theory and parameters of machine learning algorithms used by Scikit learn
- Avoid or have no access to other parts of the actual project.
It may be applicable to some people, but for many people, it can be daunting, and after a long process of preparation, learning theories and algorithms can begin the actual project. Unfortunately, this is the way most of the books on the market are being adopted.
1.2 Python machine learning
This book mainly introduces a sub-area of machine learning called predictive models. This is the most useful application of machine learning in industry. The organization of this book is mainly three kinds:
- Lesson: Learn how a child task of a machine learning project is implemented in Python, as well as best practices for this task.
- Project: Put all subtasks in tandem with actual projects for specific predictive model problems.
- Code snippet: show some Python code snippets so you can copy and paste directly for the new project.
1.2.1 Course
You need to know how to use the Python ecosystem to accomplish every sub-task in machine learning. Once you know how to use this platform to complete any of them, and get a reliable result, you can repeat the process in future projects. Let's start with the general flow of a machine learning project. A predictive model machine learning project can be divided into six major tasks:
- Defining the problem: investigating and characterization issues in order to better understand the objectives of the project.
- Analyze data: Use descriptive statistics and visualizations to better understand the data you have
- Prepare the data: Use the data conversion to better expose the structure of the predictive problem to the modeling algorithm.
- Evaluation algorithm: Design a test suite to evaluate various standard algorithms, and choose the best algorithm for the next study.
- Improved results: Use the algorithm's tuning and lifting methods to get the best performance on your data set.
- Show results: Identify models, implement predictions, and show results.
The advantage is that there are so many techniques and ways to do the same thing with this platform. In the second part you will find a simple or best practice to accomplish every subtask of a generic machine learning project. Here's a summary of the second part and a sub-task you can learn
- Lesson one: The Python ecosystem for machine learning
- Lesson two: Python and scipy 10-minute course
- Lesson Three: Importing datasets from CSV
- Lesson four: Describing data using descriptive statistical methods-analyzing data
- Lesson five: Using visualizations to understand data-analyzing data
- Lesson SIX: Preprocessing data-preparing data.
- Seventh lesson: Feature Selection-Preparing data
- Eighth lesson: Reorganization Method-Evaluation algorithm
- Lesson Nineth: Measurement-evaluation algorithm for algorithm evaluation
- Tenth lesson: Sampling Classification Algorithm-evaluation algorithm
- 11th Lesson: Sampling Regression algorithm-evaluation algorithm
- 12th Lesson: Model Selection-Evaluation algorithm
- Lesson 13th: pipelining-Evaluation algorithms
- 14th Lesson: Lifting algorithms-Improving results
- 15th lesson: Tuning algorithm Parameters-improving results
- 16th Lesson: Determining models-Showing results
You can learn about the entire machine learning process through these courses, and each course should be completed in less than 30 minutes. This book is likely to be completed on a weekend, and you can also delve into some chapters and use this book as a reference.
1.2.2 Project
Code snippets are important, you can get these in the course above, but these are not enough for your actual application, because you need a complete project from beginning to end to understand how to use Python to make predictions for the whole model. This book uses small, easy-to-understand datasets to show you this:
- The project is small.
- Very easy to operate. Datasets do not require too much feature engineering to get better results.
- Have benchmarks. Many people have used these datasets so that you can appreciate the good algorithm and understand the accuracy you can expect.
You can see three items:
- Hello Wold Iris
- Return to Boston rates
- II. Classification-sonar data sets
These projects use all the steps in part two, and then you can have an understanding of the whole machine learning process. Your future project will start here.
1.2.3 Code Snippet
Code Snippets can help you learn from a beginner to a quick start and start predicting any new project quickly.
1.2.4 You can learn.
- How to work from beginning to end from a small data set to a medium-sized dataset
- How to deliver a dataset that can be accurately predicted on unknown data
- Complete all the steps with Python
- How to learn new techniques in Python
1.3 This book is not
- Not a textbook. There is no basic theory. -This needs to be added. After all, if you do not understand the basic concept, there is no way to understand what he wants to do.
- Not algorithmic book-Introduction of No algorithms. -This self has some understanding.
- Not a Python programming book.
Let's start with 1.4.
Learn machine learning Mastery with Python (1)
