Predicting the Annual Salary Costs of Finnish Companies with Machine Learning
Valtanen, Jerry (2021)
Valtanen, Jerry
2021
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2021061537540
https://urn.fi/URN:NBN:fi-fe2021061537540
Tiivistelmä
The revenues of the Finnish Mutual Pension Insurance Companies are determined by the annual salary costs of their customer companies, as the average premium of a pension insurance is projected to be 24.4% in 2021. Therefore, knowing the annual salary costs of potential customers is crucial for determining the potential revenue that the mutual pension insurance company will receive. Suomen Asiakastieto is a public database about Finnish companies and their financial information, but some companies do not share their annual salary costs with Suomen Asiakastieto. This creates issues determining the potential revenue and assigning the customer to the appropriate salespeople. The purpose of this thesis was to find the suitable machine learning models to predict the salary costs of those companies, that have not shared their financial information with Suomen Asiakastieto. Linear Regression was used as the baseline of this thesis, as it is a simple model to implement, and it is computationally inexpensive. Neural Networks, Support Vector Regression, Decision Tree and Random Forest models were compared to the Linear Regression model and their performance was estimated using cross-validation. The results suggest that machine learning models can be used in solving the problem quite accurately, but more quality and accurate data is needed for more accurate results.