Deep learning asset forecasting

Predictions that matters

Innolab proprietary software

Predicting the movements in asset prices is one of the most difficult things to do. There are so many factors involved in the prediction that make asset prices volatile and very difficult to predict with a high degree of accuracy.

With the popularity of deep learning model in the engineering field it has attracted significant research interests in the economic and finance fields.

Statistical deep learning framework

Since mid 2015 Innolab have worked on a statistical deep learning framework that gives analysis and forecasting of asset prices every day – on a 20-day horizon. The framework offers capabilities in resembling non-linear functions, dealing with noisy, non-stationary data, and discovering hidden patterns in datasets.

The output comes from an ensemble of market research robots that has been trained to capture unknown complex nonlinear characteristics of the market. The robots constantly improve their forecasting ability as time goes by and work independently without influencing each other in their forecasts. 

Our deep learning framework cover all stages of the investment process: data analysis, signal extraction, signal translation, bet sizing, portfolio construction, risk management and execution. 

It is all based 100% on artificial intelligence and deep learning with no involvement from humans and no human rules except holding periods and volatility.

Equities Indices

51

Fixed income

16

Commodities

27

Equities


34.476

As our software and strategies have evolved, so has our approach to communicating with investors. With deep learning investment, the deep learning selection process can naturally be harder to explain.

To help, we rely on detailed visualizations and animations to demonstrate the mechanics, with the goal of demystifying the approach. Investors can also gain access to our signals and can track how these have evolved over time. We can also analyze a case-by-case basis.