Using Machine Learning to Predict Sales Conditional on Bid Acceptance

Barry E. King

doi:https://doi.org/10.14445/23939125/IJEMS-V6I11P101

Research Article | Open Access

Volume 6 | Issue 11 | Year 2019 | Article Id. IJEMS-V6I11P101 | DOI : https://doi.org/10.14445/23939125/IJEMS-V6I11P101

Using Machine Learning to Predict Sales Conditional on Bid Acceptance

Barry E. King, Jason Davidson

Citation :

Barry E. King, Jason Davidson, "Using Machine Learning to Predict Sales Conditional on Bid Acceptance," International Journal of Economics and Management Studies, vol. 6, no. 11, pp. 1-3, 2019. Crossref, https://doi.org/10.14445/23939125/IJEMS-V6I11P101

Abstract

A North American provider of vehicle parking solutions seeks to predict if a bid will be successful and, for those that are successful, what will be the cumulative sales revenue. Both traditional statistical methods and machine learning algorithms were employed. The machine learning techniques performed better than the statistical methods. There is no statistically significant difference between random forest and extreme gradient boosting for either the binary classification task or the regression task.

Keywords

logistic regression, linear regression, random forest, extreme gradient boosting, Tukey honestly significant test

References

[1] G. Janson, What Machine Learning Approaches Have Won Most Kaggle Competitions?, Retrieved from https://www.quora.com/What-machine-learning-approacheshave-won-most-Kaggle-competitions.