Profiling and Segmenting Clients with the Use of Machine Learning Algorithms
Artykuł w czasopiśmie
MNiSW
100
Lista 2021
| Status: | |
| Autorzy: | Rymarczyk Paweł, Gołąbek Piotr, Skrzypek-Ahmed Sylwia, Rzemieniak Magdalena |
| Dyscypliny: | |
| Aby zobaczyć szczegóły należy się zalogować. | |
| Rok wydania: | 2021 |
| Wersja dokumentu: | Drukowana | Elektroniczna |
| Język: | angielski |
| Numer czasopisma: | Special Issue 2 |
| Wolumen/Tom: | 24 |
| Strony: | 513 - 522 |
| Efekt badań statutowych | NIE |
| Materiał konferencyjny: | NIE |
| Publikacja OA: | TAK |
| Licencja: | |
| Sposób udostępnienia: | Otwarte czasopismo |
| Wersja tekstu: | Ostateczna wersja opublikowana |
| Czas opublikowania: | W momencie opublikowania |
| Data opublikowania w OA: | 30 czerwca 2021 |
| Abstrakty: | angielski |
| Purpose: The aim of the article is to develop a solution for customer profiling and segmentation using modern machine learning methods. Design/Methodology/Approach: Models were developed to improve the analysis of data, human behavior, data mining business processes, and as a result, the creation and provision of new improved solutions using machine learning algorithms. The GRU method was used, which is a simplified but also a more streamlined version of the LSTM cell offering similar performance with a much lower computation time. Findings: The main purpose of the developed solution is to enable and improve the analysis of profiling and segmentation of customers for forecasting sales, due to the possibility of detecting or determining additional seasonal effects. Practical Implications: Effective tools have been developed to enable customer segmentation. A more complex model was used, taking into account the sale, especially in the sense of the time series in which the sale took place. In its form, the model consists of a trend function modeling non-periodic changes in the value of time series periodic changes. Originality/Value: A novelty is the use of the GRU network, which is an improved version of the standard recursive neural network and a simplified version of the standard LSTM network. Similarly to LSTM networks, it aims to solve the problem of a vanishing gradient, i.e., its disappearance or explosion. In the presented solution, a more complex model was used, consisting of several components and taking into account sales, especially in the sense of the time series in which the sale took place. |
