Generating In-store Customer Journeys from Scratch with GPT Architectures
This addresses the need for simulating customer behaviors in retail settings, but it is incremental as it applies existing GPT architectures to a new domain.
The paper tackled the problem of generating customer trajectories and purchasing behaviors in retail stores by training a GPT-2 architecture from scratch using trajectory, layout, and scanner data, achieving more accurate reproductions than LSTM and SVM models, with fine-tuning reducing required training data.
We propose a method that can generate customer trajectories and purchasing behaviors in retail stores simultaneously using Transformer-based deep learning structure. Utilizing customer trajectory data, layout diagrams, and retail scanner data obtained from a retail store, we trained a GPT-2 architecture from scratch to generate indoor trajectories and purchase actions. Additionally, we explored the effectiveness of fine-tuning the pre-trained model with data from another store. Results demonstrate that our method reproduces in-store trajectories and purchase behaviors more accurately than LSTM and SVM models, with fine-tuning significantly reducing the required training data.