Customer Behaviour Forecast Using Deep Learning Networks

2022 7th International Conference on Machine Learning Technologies (ICMLT 2022) 2022
Agris Ņikitenko, Ilze Andersone, Andrejs Zujevs, Elīna Gaile-Sarkane, Valdis Bergs

The paper presents a case study on customer behaviour forecast within a parking products domain. In particular, authors compare different LSTM-based DL networks with and without data preprocessing to decide which of the selected architectures and hyperparameter combinations provide the least square error estimate. Unfortunately, well-known forecast methods like regression and ARIMA did not deliver the needs forecast reliability, which brought the authors to the application of DL models. While the paper does not offer novel DL models or architectures, it provides a convenient application insight enabling a better understanding of the application of the selected models. Data has been collected for several months and offers a good study backbone

Keywords
Artificial intelligence, Machine learning
DOI
10.1145/3529399.3529431
Hyperlink
https://dl.acm.org/doi/10.1145/3529399.3529431

Ņikitenko, A., Andersone, I., Zujevs, A., Gaile-Sarkane, E., Bergs, V. Customer Behaviour Forecast Using Deep Learning Networks. In: 2022 7th International Conference on Machine Learning Technologies (ICMLT 2022), Sweden, Stoholma, 1-2 March, 2022. New York: Association for Computing Machinery, 2022, pp.199-208. ISBN 9781450395748. Available from: doi:10.1145/3529399.3529431

Publication language
English (en)