Mining Customs Data for Customer Segmentation: the Case of Ethiopian Revenues and Customs Authority (Erca)
| dc.contributor.advisor | Assefa, Temtim (PhD) | |
| dc.contributor.author | Tenaw, Sewagegn | |
| dc.date.accessioned | 2019-09-26T09:53:28Z | |
| dc.date.accessioned | 2023-11-18T12:44:32Z | |
| dc.date.available | 2019-09-26T09:53:28Z | |
| dc.date.available | 2023-11-18T12:44:32Z | |
| dc.date.issued | 2018-06-04 | |
| dc.description.abstract | Customer relationship management (CRM) is the overall process of exploiting customer data and using it to increase the revenue generated from an existing customer and attract new customers by creating good relationship with them accordingly. CRM links business processes across the supply chain from back office functions through all touch points, enabling continuity and consistency across customer relationships. On the other hand, customer segmentation is the process of dividing customers into homogeneous groups, where customers within each group are similar to each other and share common attributes. In order to analyze CRM sophisticated data, one needs to explore the data by using different aspects. Data mining is one the newly emerged technology in this endeavor. Data mining finds and extracts hidden knowledge in corporate data warehouses. In this study the applicability of clustering data mining technique to support CRM activities for ERCA has been explored using the CRISP-DM process model approach. After understanding business objective of the authority, different characteristics of the ERCA customers' data were collected from the ERCA's ASYCUDA++ database. Once the customers' data were collected, the necessary data preparation steps were conducted and finally the dataset consisting of 65535 records/instances were employed to develop a clustering model. To segment customers, the k-means clustering algorithm was employed. During the cluster modeling, different experiments were conducted using different cluster numbers (k=4, 5, 6) with different seed values (10, 100, 1000). Then, the one which performed the best was selected. Thus, the model where k=4 and seed size 100 had shown better clustering performance. It enables to cluster the authority's customers into dissimilar clusters of high, medium and low value customer groups with minimum iteration value of 6 and minimum sum of square error (SSE) within clusters value of 2142.82. The results of this study have shown that the data mining techniques are valuable for customer segmentation. Hence future research directions are pointed out to come up with an applicable other data mining technique | en_US |
| dc.identifier.uri | http://etd.aau.edu.et/handle/12345678/19194 | |
| dc.language.iso | en | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | Mining Customs | en_US |
| dc.subject | Data for Customer | en_US |
| dc.subject | Segmentation | en_US |
| dc.subject | Case of Ethiopian | en_US |
| dc.subject | Revenues | en_US |
| dc.subject | Customs Authority (Erca) | en_US |
| dc.title | Mining Customs Data for Customer Segmentation: the Case of Ethiopian Revenues and Customs Authority (Erca) | en_US |
| dc.type | Thesis | en_US |