Prediction of minimum wages for countries with random forests and neural networks

Abstract

Minimum wages reflect and relate to many economic indexes and factors, and therefore is of importance to mark the developmental stage of a country. Among the 195 countries in the world, a handful of them do not have a regulated minimum wage mandated by their governments. People debate as to the advantages and disadvantages of imposing a mandatory minimum wage. It is of interest to predict what these minimum wages should be for the selected nations with none. To predict the minimum wages, motivations vary with the specific country. For example, many of these nations are members of the European Union, and there has been pressure from this organization to impose a mandatory minimum wage. Open and publicly available data from Excel Geography are employed to predict the minimum wages. We utilize many different models to predict minimum wages, and the random forest and neural network methods perform the best in terms of their validation mean squared errors. Both of these methods are nonlinear, which indicates that the relationship between the features and minimum wage exhibits some nonlinearity trends that are captured in these methods. For the method of random forests, we also compute 95% confidence intervals on each prediction to show the confidence range for the estimation.

Keywords

• random forests
• neural networks
• deep learning
• minimum wages
• prediction
• excel geography data