Anaconda: The first choice for beginner Python, entry machine learning

Now that you have learned the algorithms used in the machine learning process, how do you practice it?

Anaconda is the first choice for beginner Python and entry machine learning. It is a Python distribution for scientific computing that provides package management and environment management capabilities to easily handle multi-version python coexistence, switching, and various third-party package installation issues.

Integration package features:

NumPy: Provides the functionality of matrix operations, which are generally used with Scipy, matplotlib, the basis of all higher-level tools created by Python, and do not provide advanced data analysis capabilities.

Scipy: Depends on NumPy, which provides convenient and fast N-dimensional vector array operations. Modules are provided for optimization, linear algebra, integration, and other common tasks in data science.

Pandas: A NumPy-based tool created to solve data analysis tasks, including advanced data structures, and tools to make data analysis fast and simple

Matplotlib: Python's most famous painting gallery

Among them, Scikit-Learn is an open source machine learning toolkit integrated in Anaconda, which mainly covers classification, regression and clustering algorithms, and can directly call traditional machine learning algorithms. At the same time, Anaconda is also compatible with TensorFlow, the second generation of artificial intelligence system developed by Google, for deep learning development.

Finally, through a Python-based decision tree case, you can visually understand the process of machine learning.

A decision tree for loan applications to classify future loan applications.

The specific implementation process:

(1) Prepare the data set: select the characteristics that have the ability to classify the training data from the loan application sample data table.

(2) Build tree: select the feature with the largest information gain as the split feature to construct the decision tree

(3) Data visualization: Visualize data using Matplotlib

(4) Execution classification: used for classification of actual data. For example, enter the test data [0,1], which means there is no house, but there is work, the classification result is "household loan".

Summary

I hope that through this introduction, everyone can have a preliminary understanding of the algorithms involved in machine learning, and will help all the research on related artificial intelligence in the future. Machine learning is actually not so complicated, just applying mathematical knowledge such as statistics and probability theory to the field of artificial intelligence. With the existing software platform, it is easy to call the already integrated algorithms, so that engineers no longer worry about complex algorithm theory, and have more time to develop and innovate!

Featured question:

Question 1: Is it necessary to learn Python in order to learn machine learning?

A: Learning Python is not required, you can use other languages, but now it is more popular to use Python. Like Google's Tensorflow is to support multiple languages.

Question 2: The teacher is good. Are there open source frameworks that have ready-to-run examples (including input data, visualized output results)?

A: In EAWorld, reply to AI, you can download the code of my Python-based related case, you can run it directly to see the result.

Question 3: Is there a chance that the data foundation is poor?

A: It doesn't matter. In fact, we only need to know what the algorithm scenarios are. When we encounter similar scenes in the future, we know which algorithm to choose, and we don't need to know the specific mathematical principles of the algorithm. It's like we use the calculator every day, but we don't need to know how the computer works. And now the ready-made algorithm statements are embedded in the software and can be used directly.

Question 4: The teacher is good. If you want to systematically master the knowledge of the field of artificial intelligence from 0, there are no systematic books to recommend. Thank you!

A: "The Machine Learning" Peter Harrington (author) This book I think is a systematic introduction to the machine learning algorithm, each algorithm gives the corresponding case and code implementation.