top of page

Does artificial intelligence provide good forecasts of CAC 40 prices ?

Dernière mise à jour : 8 mars 2022

The CAC40, or CAC index, is the main stock market index on the Paris stock exchange. Its name, CAC, stands for “Cotation Assistée en Continu” (Continuous

Assisted Rating) because it is a set of values that is updated every fifteen seconds. These changes are made every working day from 9am to 5.30pm.

It represents the value of the forty largest French companies (the company's value being determined by the volume of share trading carried out).


French version :

FINETUDES.IA.FRANCAIS
.pdf
Télécharger PDF • 1.46MB

English version :

FINETUDES.IA.ENGLISH
.pdf
Télécharger PDF • 1.41MB

In more concrete terms, the CAC index reflects the general trend of the French economy. In an international context, it is not only the national economy that is represented but also the state of the world market.

As France represents around twenty percent of the European economy, the

CAC40 has a major international stake for investors. Euronext, the scientific council that is responsible for determining the stocks represented in the index, is

made up of experts in stock market operations. It is the independence of this council that allows us to anticipate the evolution of this index.

It is therefore the job of the investor that inspires our problem. Being able to predict the CAC40 share price enables us to generate profits thanks to the specific action levers of this profession.

Depending on the amount of money invested, profits can be significant.

We have limited our problem to the prediction of the evolution of the CAC index.

The study will initially consist of a state of the art of prediction technics currently used to predict stock market values by financial professionals. Secondly, we will detail the choices of artificial intelligence algorithms that we have had to use and implement.

We have implemented several models such as the LSTM model, ARIMA or GAN.

This study retraces our method and the results obtained according to the models used.

bottom of page