Development of demand forecasting model for inventory management using deep learning approach via transfer learning

Inventory Management is crucial for small and medium-sized enterprises since it requires significant financial and human resources. However, businesses now must contend with seasonal client demand in addition to a competitive and unstable economic environment. Stock predicting management inventory w...

Full description

Bibliographic Details
Main Author: Tan, Wei Qing
Format: Undergraduates Project Papers
Language:English
Published: 2023
Subjects:
Online Access:https://umpir.ump.edu.my/id/eprint/45583/
_version_ 1848827456913408000
author Tan, Wei Qing
author_facet Tan, Wei Qing
author_sort Tan, Wei Qing
building UMP Institutional Repository
collection Online Access
description Inventory Management is crucial for small and medium-sized enterprises since it requires significant financial and human resources. However, businesses now must contend with seasonal client demand in addition to a competitive and unstable economic environment. Stock predicting management inventory would be required in this situation in order to reduce losses and boost profitability. Predicting or projecting a future occurrence or trend is known as demand forecasting. Demand forecasting of inventory management is able to assist businesses from preventing overstock or stock-outs. Due to the capital commitment made by excess inventory, high inventory levels might result in revenue losses. Losses in sales and a loss in consumer satisfaction and brand loyalty may result from shortages or out-of-stock situations. To estimate future seasonal product demand, inventory management has thus integrated the traditional time series approach, machine learning, and deep learning techniques. This research discusses how to construct a demand forecasting model for inventory management via transfer learning and determine whether classical time series analysis or machine learning methods are more suitable in forecasting demand of inventory management or deep learning method is better. The models are evaluated by using confusion matrix, root mean square error (error terms) and accuracy of each model. In this research, a demand forecasting model for inventory management using transfer learning is constructed and achieved accuracy of 90.97% and error term (MAE) is 0.318. However, after the comparison between deep learning model and machine learning model as well as time series methods, LSTM model seems achieved a better result which accuracy is 91.62 and error term (MAE) is 0.316.
first_indexed 2025-11-15T04:01:01Z
format Undergraduates Project Papers
id ump-45583
institution Universiti Malaysia Pahang
institution_category Local University
language English
last_indexed 2025-11-15T04:01:01Z
publishDate 2023
recordtype eprints
repository_type Digital Repository
spelling ump-455832025-10-02T07:16:30Z https://umpir.ump.edu.my/id/eprint/45583/ Development of demand forecasting model for inventory management using deep learning approach via transfer learning Tan, Wei Qing Q Science (General) QA Mathematics Inventory Management is crucial for small and medium-sized enterprises since it requires significant financial and human resources. However, businesses now must contend with seasonal client demand in addition to a competitive and unstable economic environment. Stock predicting management inventory would be required in this situation in order to reduce losses and boost profitability. Predicting or projecting a future occurrence or trend is known as demand forecasting. Demand forecasting of inventory management is able to assist businesses from preventing overstock or stock-outs. Due to the capital commitment made by excess inventory, high inventory levels might result in revenue losses. Losses in sales and a loss in consumer satisfaction and brand loyalty may result from shortages or out-of-stock situations. To estimate future seasonal product demand, inventory management has thus integrated the traditional time series approach, machine learning, and deep learning techniques. This research discusses how to construct a demand forecasting model for inventory management via transfer learning and determine whether classical time series analysis or machine learning methods are more suitable in forecasting demand of inventory management or deep learning method is better. The models are evaluated by using confusion matrix, root mean square error (error terms) and accuracy of each model. In this research, a demand forecasting model for inventory management using transfer learning is constructed and achieved accuracy of 90.97% and error term (MAE) is 0.318. However, after the comparison between deep learning model and machine learning model as well as time series methods, LSTM model seems achieved a better result which accuracy is 91.62 and error term (MAE) is 0.316. 2023-07 Undergraduates Project Papers NonPeerReviewed pdf en https://umpir.ump.edu.my/id/eprint/45583/1/Development%20of%20demand%20forecasting%20model%20for%20inventory%20management%20using%20deep%20learning%20approach%20via%20transfer%20learning.pdf Tan, Wei Qing (2023) Development of demand forecasting model for inventory management using deep learning approach via transfer learning. Centre for Mathematical Sciences, Universti Malaysia Pahang Al-Sultan Abdullah.
spellingShingle Q Science (General)
QA Mathematics
Tan, Wei Qing
Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title_full Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title_fullStr Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title_full_unstemmed Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title_short Development of demand forecasting model for inventory management using deep learning approach via transfer learning
title_sort development of demand forecasting model for inventory management using deep learning approach via transfer learning
topic Q Science (General)
QA Mathematics
url https://umpir.ump.edu.my/id/eprint/45583/