Many people like python not only because it is easy to learn and easy to get started, but also because python has powerful third-party libraries. So in machine learning, which libraries can python be used for?
1. Scikit-learn: One of the most popular ML libraries, supporting many supervised learning and unsupervised learning algorithms. Based on two python libraries, Numpy and Scipy, it provides a set of algorithms for common machine learning and data mining, clustering, regression and classification.
2. Tensorflow: When you use python to write code, you can compile and run it on your CPU or GPU. You don't need to write C++ or CUDA code, you can run it on a GPU cluster.
3. Theano: Another excellent class library for numerical calculation, similar to Numpy. Theano allows you to efficiently define, optimize and evaluate mathematical expressions involving multidimensional arrays. The difference is that it uses the computer's GPU to perform data-intensive calculations 100 times faster than when running on the CPU alone.
4. Pandas: A very popular library that provides easy-to-use and intuitive and efficient data structures. There are many built-in methods to group, combine data and filter, and perform time series analysis. You can easily obtain data from different sources such as sql database, CSV, Excel, and JSON files, and perform operations on the data.
5. Seaborn: A popular visualization library based on Matplotlib. It is an advanced library, which means it is easier to generate certain types of graphs, including heat maps, time series, etc.