QuantDesk^® Stock Forecaster

Identify investment opportunities ahead of the broad market

What is the Forecaster?

The QuantDesk® Forecaster predicts future asset prices using multi-factor models employed by state of the art Machine Learning algorithms. The forecaster works in a two-stage process:

First, a genetic algorithm (GA) is employed to discover a sub group of 10 to 15 out of a total of 950 indicators that are most predictive when employed together.
Next, the selected factors are used to build a regression model that predicts future benchmark-relative change in price along with corresponding confidence measures.
Users can construct their own multi-factor model based on their own proprietary research and insight. Lucena’s large dataset of factors define the state of each security over time.

How the Forecaster Works

Choose the universe of assets to forecast:

You can choose a portfolio, an index of stocks, a scanned list or a canned list to forecast its constituents.

Choose your indicator-set model, lookback and forecast period:

Select Lucena’s default model, based on the machine’s determination of the most predictive factors, or build your own model. The forecast will then “train” (study repeatable patterns) over the lookback period and will assess its price velocity and alpha decay into the future for the forecasted period.

Click forecast and inspect the results:

QuantDesk® provides a tabular and graphical summary of forecast results. The forecast for each security includes a predicted price change, a confidence score, as well as upper and lower variance bounds.

Why use the Forecaster?

Construct your own custom model.

Enhance indicator-set selection by consulting the machine to discover the most predictive set of indicators for your security basket.

Assess the predictive power of the Forecaster by traveling back in time and simulating trades derived from the forecaster output.

Discover more:

QUANTDESK® MODULES

Interested in taking the next step?

LET’S TALK

QuantDesk® Stock Forecaster