As I’ve mentioned in previous posts, I am a big fan of Kaggle. The notebook below is my final submission to the RentHop: Rental Listing Inquiries competition which ended a few months ago. This relatively simple submission placed in the top 64%.
In the following notebook, I will walk you through the machine learning process as it applies to this competition. You will learn to extract some basic features from more complex features like the date or the description. You will also see how to engineer some new simple features with basic math.
Continue reading Machine Learning Projects – A Guide to Predicting Popularity of Rental Listings on RentHop
In this notebook, I show my process for developing an automated trading system that I currently am trading live. Because I am currently trading this system I have chosen to not release the full code and data set that I have used.
However, in this notebook you will learn to:
- Follow an efficient process when solving a machine learning problem
- Find the most important features that drive your outcome label
- Determine the proper way to frame a trading problem
- Understand how to get trading results from a modeled system
Continue reading Machine Learning Projects – How I Developed an Automated Trading System Using Machine Learning
I’ve spent over six years as a full-time equity trader. Two of these years I’ve spent developing automated trading strategies. In this notebook here I show one way of developing a machine learning model to make trading decisions.
This model uses the EOD-Stock-Prices data set that is freely available at Quandl. Using only end-of-day price and volume I add some basic technical features to the data set and then train a RandomForest model to predict potential returns.
Continue reading Machine Learning Projects – Building a Simple Stock Trading System using Quandl Data
To get good at machine learning and data science, you need to practice a lot. One of the best ways to get real world experience and raise your publicity as a machine learning practitioner is to compete in machine learning competitions.
Kaggle is probably the biggest and most popular machine learning competition website. I am a big fan of it. The competitions are well organized and the community is amazingly helpful.
Continue reading Machine Learning Projects – A Guide to your First Kaggle Machine Learning Competition
As is common when learning a new language, the ‘Hello World’ problem is a rite of passage that takes you from start to finish with the most basic of applications. Building this end to end project is a vital first step before learning to do anything more advanced.
This problem uses the Iris data set which is a small, easy to work with data set that requires minimal preparation before modeling. This is also a great example of following the machine learning process step by step.
Continue reading Machine Learning Projects – The Hello World Problem