In our previous article “ChatGPT: boarding for better and worse”, we saw that this conversational agent developed by Open AI is a tool of formidable efficiency. “You can ask ChatGPT for almost anything: writing a poem, a speech, an article, a song, in almost any language. It can also help write codes and programs that can be used by cybercriminals and online scammers to generate scam and phishing emails and thus increase threats such as ransomware. And could it provide quantitative forecasts from a history of a variable whose forecast we are looking for?
This is the subject of this post
Few examples at the moment, but yes ChatGPT can produce a quantitative forecast
In an article titled “Creating a stock price forecasting system with Chat GPT”, the author of the post wanted to build a prediction system of stock prices, and he wanted to test ChatGPT to find out if it is capable of such a move and how the predictions might perform.
To do this, he submitted the following question to the chatbot: “How to create a Python script that makes a prediction about the future price of a stock? “and asked for a sample script. And here is the script he got:
While examining the result, the author noticed that the script loaded a file that it had to generate. He also noticed that the script did retrieve data sets. He knew with Python how to organize this dataset, but the generated script alone would not work.
He also noticed that the script used a linear model for the prediction. But he was very skeptical of the predictions given the structure of the data which did not lend itself to such a model.
Interesting, but not satisfying
He submitted another request to ChatGPT to get the same script with a neural network model:
As can be seen in the second script, the code is very similar but it uses the neural network model instead of the linear regression model.
In conclusion, the author remarks that the scripts that ChatGPT generated are quite good. They allow to generate a good structure to work on, although he considers that the person who implements the script must have the necessary knowledge to interpret the generated code.