Company Classification Using Machine Learning Models

The current information age is flooded with data, much of which is complicated and difficult to interpret. It has therefore become necessary to create computational tools that allow for the processing and analysis of large datasets. Main goal of this contribution is to examine the effectiveness of fiveteen different machine learning algorithms for classifying a business from the Forbes 2000 ranking of the world's largest companies, based on their market value statistics. Used algorithms cover also simple models such as Logistic Regression, Naive Bayes, k-nearest neighbour, Generalized Linear Models and computationally challenging models such as tree based algorithms, ensemble methods and neural networks. Contribution shows that k-nearest neighbour algorithm and ensemble methods give better results of classifying banking and non-banking companies based on market value statistics than traditional tree based methods or neural networks for this dataset in terms of classification accuracy, sensitivity and specificity.