Python is an object-oriented interpreted computer programming language with rich and powerful libraries, plus its simplicity, ease of learning, fast speed, open source free, portability, extensibility and object-oriented features, Python has become a The most popular programming languages of 2017!
Artificial intelligence is one of the hottest topics at the moment. Machine learning technology is an essential skill for artificial intelligence implementation. The Python programming language contains the most useful machine learning tools and libraries. The following are the top ten machine learning libraries that Python development engineers must know!
一、Scikit-Learn
In machine learning and data mining applications, Scikit-Learn is a powerful Python package that we can use for classification, feature selection, feature extraction and aggregation.
、 、 State models
Statsmodels is another powerful library focused on statistical models, mainly for predictive and exploratory analysis, fitting linear models, performing statistical analysis or predictive modeling, using Statsmodels is very suitable.
3. PyMC
PyMC is a tool for making "Bayesian curves", which includes diagnostic tools for Bayesian models, statistical distributions, and model convergence, as well as some hierarchical models.
4. Gensim
Known as "people's topic modeling tool", Gensim focuses on Dirichlet partitions and variants, which supports natural language processing, making it easier to combine NLP and other machine learning algorithms, and also cites Google's based on Text representation word2vec for recurrent neural networks.
5. Orange
Orange is a library with a graphical user interface that is fairly complete in classification, aggregation, and feature selection methods, as well as cross-validation methods.
6. PyMVPA
PyMVPA is a statistical learning library that includes cross-validation and diagnostic tools, but is not as comprehensive as Scikit-learn.
7. Theano
Theano is the most mature deep learning library. It provides a good data structure to represent the layers of the neural network. It is very efficient for linear algebra. Similar to Numpy's arrays, many Theano-based libraries are using its data structure. It also Supports GPU programming available out of the box.
Eight, PyLearn
PyLearn is a Theano-based library that introduces modularity and configurability to Theano to create neural networks with different configuration files.
9. Hebel
Hebel is a neural network library with GPU support, which can determine the properties of neural networks through YAML files, provides a friendly way to separate god-level networks from code, and runs models quickly. It is written in pure Python and is very Friendly library, but lacking in depth and breadth due to recent development!
10. Neurolab
Neurolab is an API-friendly neural network library that contains different variants of recurrent neural network implementations, and if using RNNs, this library is one of the best APIs of its kind.
The above are the top ten machine learning libraries that Python development engineers must know. In addition, there are machine learning libraries such as OverFeat, Nolearn, and Decaf. I will not introduce them one by one here. If you are interested, you can learn more!
It is recommended to take a look at the courses of Old Boys Education . Since Python development started in 2012, the curriculum system has matured, and there are currently full-time Python full-stack development classes with 0 basics that can be learned, as well as on-the-job operation and maintenance, testing, front-end development and other friends We can learn the Python Automation Development Weekend Course. Now the Python full-stack development has begun to implement the "dual-track 5+5" learning mode. The course adds artificial intelligence (image recognition, drones, unmanned vehicles, etc.) and related knowledge of the go language to expand everyone's knowledge field.