Member-only story
Temporal Fusion Transformers
A Deep Learning Strategy Applied to financial asset
@panData26 min read·Just now--
I’m now bringing you a project where we can predict investment returns using a cutting-edge architecture: Temporal Fusion Transformers.
Time series modeling is a science in itself. What I intend here is to apply Deep Learning — using the most modern architectures available today — to model and predict the behavior of a time series.
In this case, we want to predict investment returns on financial assets in the luxury fashion sector. I decided we’re going to analyze LVMH (Moët Hennessy Louis Vuitton), the world’s largest luxury conglomerate.
Being a global company, it’s traded on different stock exchanges under different ticker codes. On Euronext Paris, for example, it trades under MC.PA — and that's exactly the ticker we'll use to download the stock data.
LVMH owns 75 brands (called “Maisons”), organized into 6 main divisions — a portfolio that ranges from expensive wines to ultra-luxury hotels.
The reason I chose this company? I’m here in Florence, and I see countless job opportunities in these major companies. So why not take a closer look at all that potential, right?
Before we get into it — this is a didactic project. Nothing here is investment advice or…
