Financial Forecasting, Planning and Analysis Using Machine Learning

Abstract

This paper studies multiple linear regression and time series regression: two regression methods, the task of which is to learn the relationship between various components of financial statements and the profits earned. Time series regression is used to understand and predict the behavior of dynamic systems such as the modeling and forecasting of economic, financial, biological, and engineering systems.

For the machine learning methods that have been used to conduct this study, we have used a dataset that was created by using information provided by the luxury brands themselves to maintain accuracy and authenticity of data. This will help in providing and gaining accurate information.

Three machine learning models; simple linear regression, multiple linear regression and times series regression were employed for conducting this study where we try to predict future values based on the historical data provided to us. The time series regression model then has various models which have been used after comparing and determining the best suited model according to the component that we have tried to predict.

Our study adds to the existing literatures by providing insights into which models are better suited for the type of predictions and the components when related to each other and time. Specifically, we find that all the brands in the fashion industry had their lowest financial performance in the year 2020 when compared with the data of past 10 years, but have also managed to recuperate and flourish since.

The findings of the study have important implications for investors and fashion enthusiasts who want to explore different investment opportunities other than merchandise purchasing. These individuals can use the results given to make investments decisions which can be worth a large sum in monetary terms.