Machine Learning Model: Python Sklearn & Keras
- English
- Artificial Intelligence
Machine Learning
- Project length: 17h 18m
Machine learning (ML) is one of the liveliest fields of artificial intelligence (AI) that is causing major disruptions in our world today. And, to sink your feet deeper in the field, you need to know how to build machine learning models that can accurately predict outcomes. Two of the most popular Python libraries for building machine learning models are Scikit-learn and Keras. In this hands-on project tutorial, you’ll learn how to work with the two libraries in building amazing models for solving various machine learning problems. This is the practical project you need to take your machine learning career to the next level.
Overview
Introduction
In this project tutorial, we are going to create two simple machine learning applications in Python 3.5+ using two totally different libraries: Scikit-learn and Keras. We’ll cover all the details, including resources, tools, and languages, that are necessary to build a Python machine learning model—starting from the basics such as setting up the right tools and frameworks to more advanced topics related to the development. In this classification problem, our model will be trained to classify the dataset from the open Hedge Fund, which can be accessed at www.numer.ai. The dataset consists of 21 features and 2 categories or classes. This problem is complex because the data we’ll be trying to predict comes from the volatile stock market data. Before diving into the main task, we’ll see how a “Hello World” in machine learning looks like. After that, we’ll learn how to use numer.ai and their pre-defined datasets. After brute forcing the framework and method parameters, you’ll gain enough skills to create your own machine learning model.
What are the requirements? To speed up your learning in this project, you need the following:
- Basic skills in Python 3.5+
- Basic working knowledge of Jupyter Notebook, PyCharm ID, or Spyder3 IDE
- Stable Internet connection
- A lot of curiosity for machine learning
Who is the target audience?
- Do you want to start a career in the exciting world of machine learning?
- Do you want to build simple machine learning models?
- Do you want to experience how to use Scikit-learn and Keras machine learning libraries?
- Do you want to know the parameters of a ML-model you can tweak to get optimized results?
What will you learn after finishing this project?
- How to create your own machine learning model for predicting outcomes
- How to use the Scikit-learn and Keras Python-based machine learning libraries
- How to use Python for authoring simple machine learning models
- How to create and apply the best machine learning practices for your specific use cases
When are the streaming sessions (streaming schedule)?
Weekly, 3:00 pm EST on Saturdays and Sundays
Project Outline
Session 1: ML Models: Setting up Environment (Anaconda, IDEs, etc.)
- Setting up the programming environment for Windows (Installing Python 3.5+, PyCharm, Spyder3, and Jupyter)
- Taking a look at different ML-libraries for Python
- Going through various methods and algorithms present in Scikit-learn, Keras, and Pandas
Session 2: Cheatsheets and different Machine Learning Problems
- Python 3.5+
- Scikit-learn, Keras, Pandas, and Numpy
- First Steps with Pandas and Numpy; visualizing and exploring our datasets
- Visualizing data with Orange
Session 3: Features, Targets, Pandas and Data
- Use our labeled data to build an ML model
- Get the accuracy of the model
- Optimize its parameters for better results
Session 4: Data Mining
- What is Numer.ai and how does it work
- What does the dataset represent?
- Explorative analysis and visualization of the dataset
Session 5: Hello World in Machine Learning Part
Session 5.1: Hello World in Machine Learning Part 1/3
- Most common and basic ("Hello World") example in the field on machine learning
Session 5.2: Hello World in Machine Learning Part 2/3
- Using a UI Tool "Orange" to perform Machine Learning tasks with a simple Drag&Drop.
- You can perform complex Data Analytics with Orange without knowing how to code.
Session 5.3: Hello World in Machine Learning Part 3/3
- Splitting Data
- Confusion Matrix
Session 6: Introduction into the Numerai Network/Project
- Explanation of real-world Machine Learning Challenge
- The greater your ML Model Accuracy/Score, the more money you earn
Session 7: First steps into Numerai and Kaggle
- Trying to read Numerai Dataset.
Session 8: Basic Understanding of Class separation
- Explanation of Class Separation
- Visualization of class separations
Session 9: Basic Comparison between SVM and Neural Nets
- Explanation of low-level computation/equations of a neural network
- Underlying structure of a neural network (biases, neurons, weights, etc.)
Session 10: Recognizing Handwritten Digits with Neural Nets
- import MNIST Dataset
- Preparing Handwritten Digit Recognition
Session 11: Fitting Machine Learning Model in Sklearn
- Use Data to train our ML Classifier
- First insights about stock market trading
Session 12: Using Stock Exchange Data in Machine Learning
- Importing stock market data from Yahoo
- First look, and visualizing stock data
Session 13: Finalizing The Project
- Creating MLPClassifier
- Final thoughts on Machine Learning and Stock Market Trading
Tools: Orange, Jupyter. PyCharm, Anaconda