Recommendation for Product Selling Opportunity using Hybrid-MCDM in E-commerce Marketplace
The rapid growth of E-commerce platform has attracted both consumers and sellers, yet it presents significant challenges for sellers due to intensifying competition. This heightened competition may result in market losses for sellers. To mitigate these challenges, sellers must enhance their competit...
| Main Authors: | , , |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
Christopher Chee
2025
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/48239/ |
| Abstract | Abstract here |
| Summary: | The rapid growth of E-commerce platform has attracted both consumers and sellers, yet it presents significant challenges for sellers due to intensifying competition. This heightened competition may result in market losses for sellers. To mitigate these challenges, sellers must enhance their competitiveness in the marketplace. Thus, a data analytic approach to identify potential products through product selling recommendation for sellers within the E-commerce marketplace was proposed. By leveraging these recommendations, sellers can make informed decisions and saves time on complex decision-making processes. The Multi-Criteria Decision-Making (MCDM) method is applied to identify potential products, utilizing Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). AHP uses pairwise comparison to derive weights, while TOPSIS focuses on proximity to the ideal solution. These methods have been selected for ranking alternatives in MCDM. In order to apply MCDM, various product feature such as Estimated Sales Volume (ESV), Net Promoter Rating (NPR), Sales Rate (SR) and Price (P) are proposed. These features serve as the key metrics for evaluating the potential of a product in the marketplace. The hybrid-based MCDM method (AHP-TOPSIS) is evaluated using Ranking Evaluation Value (REV). REV is used as a quantitative metric to compare the appropriateness of the ranking outcomes under a consistent set of criteria weights. In this evaluation, higher REV values indicate better-aligned recommendations with the decision making objectives. To ensure the consistency of the hybrid-based ranking model, further experiment is conducted to evaluate its overall performance. To test the consistency of the model over time, different datasets were used to imitate the data from various timelines. Additionally, different product categories were included to evaluate the performance of the model across diverse types of products. The results demonstrated that AHP-TOPSIS offers superior identification of potential products based on the product features (ESV, NPR, SR and P). Therefore, the application of AHP-TOPSIS to identify potential products is able to help sellers to overcome competitiveness in the E-commerce marketplace. |
|---|