CATS: Clustering-Aggregated and Time Series for Business Customer Purchase Intention Prediction

TL;DR

The paper tackles the problem of predicting customers' next purchase under a head-to-tail distribution by introducing CAGRU, a two-stage framework that first clusters customers using k-shape clustering on multi-modal time-series data and then applies an AttentionGRU-based predictor within each cluster. This segment-specific approach aims to mitigate data imbalance and capture both short-term and group-level purchase patterns. Empirical results on four poultry marketing datasets show CAGRU consistently outperforms seven baselines across metrics such as accuracy, AUC, and F1, with the largest gains on the largest dataset. Ablation and parameter analyses demonstrate that clustering contributes substantially to performance and that optimal cluster counts depend on dataset size, indicating practical guidance for deployment in real-world, imbalanced time-series forecasting. The work offers a scalable, domain-agnostic strategy for enhancing purchase-intention forecasting in multi-modal, imbalanced settings and can extend to other multivariate time-series tasks.

Abstract

Accurately predicting customers' purchase intentions is critical to the success of a business strategy. Current researches mainly focus on analyzing the specific types of products that customers are likely to purchase in the future, little attention has been paid to the critical factor of whether customers will engage in repurchase behavior. Predicting whether a customer will make the next purchase is a classic time series forecasting task. However, in real-world purchasing behavior, customer groups typically exhibit imbalance - i.e., there are a large number of occasional buyers and a small number of loyal customers. This head-to-tail distribution makes traditional time series forecasting methods face certain limitations when dealing with such problems. To address the above challenges, this paper proposes a unified Clustering and Attention mechanism GRU model (CAGRU) that leverages multi-modal data for customer purchase intention prediction. The framework first performs customer profiling with respect to the customer characteristics and clusters the customers to delineate the different customer clusters that contain similar features. Then, the time series features of different customer clusters are extracted by GRU neural network and an attention mechanism is introduced to capture the significance of sequence locations. Furthermore, to mitigate the head-to-tail distribution of customer segments, we train the model separately for each customer segment, to adapt and capture more accurately the differences in behavioral characteristics between different customer segments, as well as the similar characteristics of the customers within the same customer segment. We constructed four datasets and conducted extensive experiments to demonstrate the superiority of the proposed CAGRU approach.

Figures (8)

• Figure 1: Distribution of the relationship between customer activity and customer quantity.
• Figure 2: Comparison of the effects of the Hamming distance before and after clustering. The distribution of the Hamming distances before clustering is shown in \ref{['fig:all customers']}, while \ref{['fig:cluster 1']}- \ref{['fig:cluster 3']} show the distributions of the three clusters after clustering.
• Figure 3: The visualization of five randomly selected sample for each group.
• Figure 4: Overview of the proposed CAGRU model
• Figure 5: Architecture diagram of the clustering module