Χρήση μεθόδων μηχανικής μάθησης στο τομέα της εφοδιαστικής αλυσίδας : πρόβλεψη Backorder
Use of machine learning methods in the supply chain sector : Backorder prediction
View/
Keywords
Supply chain ; Machine learning ; Backorder prediction ; Neural networks ; Microsoft AzureAbstract
As the amount of data that businesses have to manage has increased rapidly in recent years, a fact to which
the spread of the internet and the electronic platforms that make it up have undoubtedly contributed, the
field of machine learning and big data is becoming increasingly useful in the proper management of this
data, saving time and valuable resources in businesses.
In this particular thesis we will focus on the part of the Supply Chain and more specifically on the part
concerning Backorder Prediction, which is now a task of particular importance in medium and large
companies due to the overconsumption and also the unpredictable/changing nature of market demand.
Because the number of backorders is much smaller than the number of orders shipped on time,
implementing an effective forecasting model for this sector is a challenge.
Various machine learning and deep learning methods and models will be used and compared, to calculate
if a product will be backordered, i.e., an order that cannot be completed immediately due to a lack of
availability of the product.
The structure of the thesis begins with an automated process written in Python, where all the csv files
containing the columns with our data are received and a basic conversion and cleaning is done to them, so
that they transform into a format recognizable by SQL database. Then, the desired pre-processing of the
data is done and finally they are uploaded to Microsoft's cloud service, Azure SQL, where they can be further
managed and read. Afterwards, in order to run the machine learning models and visualize, compare and
interpret the results, we convert the tables from the database into dataframes to continue the
aforementioned processes in Jupyter Notebook.
Postgraduate Studies Programme
Πληροφοριακά Συστήματα και ΥπηρεσίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνNumber of pages
87Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα