Hybrid Model for Customer Segmentation Based on RFM Framework

Abstract

With the development of big data technologies in recent years, companies that can efficiently use customer data are gaining an edge in the market competition. For e-commerce platforms, by customer segmentation, the cost of marketing can be effectively reduced, customer satisfaction can be improved, and customer lifecycle value (CLV) can be increased. However, in the past RFM models and K-means algorithms, high-dimensional data could not be taken into consideration and the accuracy rate is low. Here we propose a new method that employs entity embedding to handle the category variables and uses the GBDT algorithm for feature extraction and then uses MLP for prediction to classify customers into eight categories. This new method can significantly improve the accuracy rate. Furthermore, this class label is based on the RFM model, which can efficiently identify customers’ value by supervised learning and provide strong support for the customer segmentation strategy of e-commerce platforms. We validate the effectiveness of the proposed approach on a real-world dataset of customer consumption information. The experimental results illustrate that our model significantly outperforms baseline models. The accuracy of proposed in this paper increased by about 20.4% relative to GBDT and about 3.8% relative to MLP.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.