Using Machine Learning Methods to Solve Problems of Forecasting Demand for New Products in the Internet Marketplace

304

Abstract

The work is aimed at researching the possibility of using machine learning methods to build models for forecasting demand for new products in the online store Ozon. ru. Approaches to the solution that were not previously used in a specific task are proposed for consideration. Data on sales history and storage of goods at Ozon.ru are used as a sample. There is a description and analysis of the approximate loss of the Ozon.ru website, the data used, the process of building a base model, and the results obtained. It describes the metrics used to evaluate the prediction results and makes a comparative analysis between the prediction results of the built model and the results of heuristically selected values.

General Information

Keywords: demand forecasting, new products, encoding, gradient busting, regression, preprocessing, data processing, machine learning

Journal rubric: Data Analysis

DOI: https://doi.org/10.17759/mda.2020100404

For citation: Osin A.A., Fomin A.K., Sologub G.B., Vinogradov V.I. Using Machine Learning Methods to Solve Problems of Forecasting Demand for New Products in the Internet Marketplace. Modelirovanie i analiz dannikh = Modelling and Data Analysis, 2020. Vol. 10, no. 4, pp. 41–50. DOI: 10.17759/mda.2020100404. (In Russ., аbstr. in Engl.)

References

  1. Bisong E. Introduction to Scikit-learn // Building Machine Learning and Deep Learning Models on Google Cloud Platform 2019. P. 215–229.
  2. Cerda P., Varoquaux G., Kégl B. Similarity encoding for learning with dirty categorical variables // Machine Learning. 2018. P. 1477–1494.
  3. Ke G. et al. Lightgbm: A highly efficient gradient boosting decision tree // Advances in neural information processing systems. 2017. P. 3146–3154.
  4. Redell N. Shapley Decomposition of R-Squared in Machine Learning Models // arXiv preprint arXiv:1908.09718. 2019.
  5. Botchkarev, Alexei. “Performance metrics (error measures) in machine learning regression, forecasting and prognostics: Properties and typology.” // arXiv preprint arXiv:1809.03006.
  6. Al Daoud E. Comparison between XGBoost, LightGBM and CatBoost Using a Home Credit Dataset // International Journal of Computer and Information Engineering. 2019. P. 6–10.

Information About the Authors

Artem A. Osin, Moscow Aviation Institute (National Research University), Moscow, Russia, ORCID: https://orcid.org/0000-0002-2664-1370, e-mail: artemosin1@yandex.ru

Artem K. Fomin, Moscow Aviation Institute (National Research University), Moscow, Russia, ORCID: https://orcid.org/0000-0003-3545-4435, e-mail: artem.fomin@outlook.com

Gleb B. Sologub, PhD in Physics and Matematics, Associate Professor of the Department of Mathematical Cybernetics of Institute of Information Technologies and Applied Mathematics, Moscow Aviation Institute (National Research University), Moscow, Russia, ORCID: https://orcid.org/0000-0002-5657-4826, e-mail: glebsologub@ya.ru

Vladimir I. Vinogradov, PhD in Physics and Matematics, Associate Professor, Moscow Aviation Institute (National Research University), Moscow, Russia, ORCID: https://orcid.org/0000-0003-3773-9653, e-mail: vvinogradov@inbox.ru

Metrics

Views

Total: 490
Previous month: 17
Current month: 22

Downloads

Total: 304
Previous month: 23
Current month: 35