This paper deals with the tricky issue of forecasting the number of daily orders received by a delivery company that operates through the internet. The research tries to address the problem through the Multilayer Perceptron Neural Network (MLP). The more important step of the methodology is the identification and characterization of the features to adopt as inputs for the MLP in the cited case. The number of visits (NV5) to the company website, months, days of the week and the public holidays are the four features used to predict the number of received orders (NOs). The Levenberg Marquardt back-propagation algorithm was used to train the model. The proposed methodology was applied by a delivery company, which operates in Italy, to forecast the daily demand. The results showed a good accuracy of the MLP in predicting the NOs, with a Root Mean Squared Error of the 20% from the actual NOs. Copyright (C) 2022 The Authors.

Demand forecasting for delivery platforms by using neural network

Caterino, M;
2022-01-01

Abstract

This paper deals with the tricky issue of forecasting the number of daily orders received by a delivery company that operates through the internet. The research tries to address the problem through the Multilayer Perceptron Neural Network (MLP). The more important step of the methodology is the identification and characterization of the features to adopt as inputs for the MLP in the cited case. The number of visits (NV5) to the company website, months, days of the week and the public holidays are the four features used to predict the number of received orders (NOs). The Levenberg Marquardt back-propagation algorithm was used to train the model. The proposed methodology was applied by a delivery company, which operates in Italy, to forecast the daily demand. The results showed a good accuracy of the MLP in predicting the NOs, with a Root Mean Squared Error of the 20% from the actual NOs. Copyright (C) 2022 The Authors.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11386/4812873
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact