At 3Fourteen, our team combines years of market experience with world-class data science. This combination ensures our models are rigorous, realistic, and cutting edge. Each core model begins with an understanding of key drivers specific to that market. From there, we analyze these indicators by matching the best statistical methodology to that time series. During this process, we draw upon a wide variety of mathematical concepts including: machine learning, pattern matching, decision trees, clustering algorithms and others. When appropriate, we utilize extensive out-of-sample testing to ensure that our parameters are robust, and that the model is not overfit to a past we are unlikely to experience going forward.
Below, we lay out a schematic of our Core Crude Oil ML Model as an example. This model synthesizes four broad categories of market indicators with a machine learning framework. Each model component is extensively trained and then tested on out-of-sample data. The result: The most rigorous and sophisticated crude oil market in the investment research space. If you have a modeling problem or question, we also provide consulting services and custom solutions.
Core Crude Oil ML Model
Key inputs synthesized in a probabilistic ML framework:
Our Key Models
Fundamentals, positioning, and pricing wrapped in an ML framework.
Real Asset Allocation
Reconstructing the 60/40 portfolio with real assets – which ones and how much.
Real rates, positioning, and pricing wrapped in an ML framework.
Relative and absolute technical model using ML.
S&P 500 Risk
Our key risk gauge for the stock market.