Machine Learning and Event Study to Explore the Influence of ChatGPT on Microsoft Stock

Abstract

In light of the increasingly vital role of Artificial Intelligence (AI) has played in this era, it is imperative to conduct a comprehensive examination of the exact impact of specific event associated with AI on advanced corporations, economies and even countries. This study focuses on analyzing the recent event involving the connection of ChatGPT to Microsoft. To be more specific, this paper employed 3 machine learning models and one Stata model event study to respectively identify the best fitted model and its precise impact. In this work, 3 machine learning have been used, namely Support Vector Regression (SVR), K-Nearest Neighbor categorization algorithm (KNN) and Random Forest, to spot the model that fits the Microsoft stock the best. Initially, data was collected from Yahoo Finance, which is set and indexed in advance. Subsequently, data is individually put to train the 3 models. Ultimately, Mean Square Error (MSE), Root Mean Squared Error (RMSE) and R squared score are calculated with care and compared to obtain the results. Additionally, after collecting the data, a sensible window of event study has been set. Experimental results demonstrated that Random Forest performs the best among the 3 models and the specific event of ChatGPT connecting to Microsoft has a limited effect on the firms stock price.