I want to learn Python for my machine learning application. Can you help me with resources like tutorial, website, book name etc. ? Also want to know which IDE will be better for machine learning application development?
The suggested method below is unorthodox but in my experience it has worked wonders with people having little to no prior knowledge in Python and ML.
Step 1. Complete the Python 3 course on SoloLearn Android/iOS app.
Rather than jumping straight into scikit-learn, pandas, numpy, tensorflow, keras etc. it is important to understand the basic concepts of a high level programming language and the app vastly speeds up the process.
Step 2. Complete Andrew Ng's Machine Learning course on Coursera.
Even though the course used MATLAB/Octave for programming excercises it is not important to complete those. The important objective is to understand the mathematics of machine learning i.e. Statistics, Probability, Calculus, Linear Algebra.
Step 3. Learn the syntax of python libraries necessary for ML
Start with numpy (Numerical Python). After the ML course, you will be able to relate the linear algebra operations of numpy to ML algorithms. Then move onto matplotlib for learning visualization techniques and norms. Learn pandas for handling data files. Last but not least, sklearn (Scikit-Learn) package which contains implementations of almost all state of the art machine learning algorithms as well as data preprocessing, scaling etc. packages.
Tip: Try to complete the assignments in ML course using Python 3 rather than MATLAB/Octave. That way you will be applying what you are learning.
IDE: After trying many options, I now reside to Google Colab, a cloud based platform totally free for research and non-commercial applications. It is a Linux virtual machine with 16 GB of RAM and 16GB of VRAM (NVIDIA Tesla GPU) (now comes with a TPU (Tensor Processing Unit) runtime as well) with Python 3 and all other required packages for ML (numpy, matplotlib, pandas, sklearn, tensorflow, keras etc.) preinstalled. Just load your data files in Google Drive, open a Google Colab notebook and start working, hassle free.
For deep learning please look into Stanford''s Convolutional Neural Network Course freely available on YouTube as well as the resources suggested by Helge Egil Seime Pettersen. TensorFlow is the leading framework for deep learning and has a Python API. The high level wrapper for TensorFlow, known as Keras, accelerates the model prototyping phase.
Hello Kazi Mahmudul Hassan , python is the best choice in terms of a programing language for machine learning.
My favorite IDE is Jupyter Lab, you can find more information here:
https://medium.com/@boyanangelov/is-this-the-best-data-science-ide-jupyter-lab-review-fdd165470f13
In regards to a tutorial, my recommendation is to start with Michael Nielsen online book:
http://neuralnetworksanddeeplearning.com/about.html
Hope this helps.
Good luck!
The suggested method below is unorthodox but in my experience it has worked wonders with people having little to no prior knowledge in Python and ML.
Step 1. Complete the Python 3 course on SoloLearn Android/iOS app.
Rather than jumping straight into scikit-learn, pandas, numpy, tensorflow, keras etc. it is important to understand the basic concepts of a high level programming language and the app vastly speeds up the process.
Step 2. Complete Andrew Ng's Machine Learning course on Coursera.
Even though the course used MATLAB/Octave for programming excercises it is not important to complete those. The important objective is to understand the mathematics of machine learning i.e. Statistics, Probability, Calculus, Linear Algebra.
Step 3. Learn the syntax of python libraries necessary for ML
Start with numpy (Numerical Python). After the ML course, you will be able to relate the linear algebra operations of numpy to ML algorithms. Then move onto matplotlib for learning visualization techniques and norms. Learn pandas for handling data files. Last but not least, sklearn (Scikit-Learn) package which contains implementations of almost all state of the art machine learning algorithms as well as data preprocessing, scaling etc. packages.
Tip: Try to complete the assignments in ML course using Python 3 rather than MATLAB/Octave. That way you will be applying what you are learning.
IDE: After trying many options, I now reside to Google Colab, a cloud based platform totally free for research and non-commercial applications. It is a Linux virtual machine with 16 GB of RAM and 16GB of VRAM (NVIDIA Tesla GPU) (now comes with a TPU (Tensor Processing Unit) runtime as well) with Python 3 and all other required packages for ML (numpy, matplotlib, pandas, sklearn, tensorflow, keras etc.) preinstalled. Just load your data files in Google Drive, open a Google Colab notebook and start working, hassle free.
For deep learning please look into Stanford''s Convolutional Neural Network Course freely available on YouTube as well as the resources suggested by Helge Egil Seime Pettersen. TensorFlow is the leading framework for deep learning and has a Python API. The high level wrapper for TensorFlow, known as Keras, accelerates the model prototyping phase.
- Badges
- Science topic