PriceMorphNet: Cross-Attention Multimodal Fusion for Intelligent E-Commerce Price Prediction

Author(s)	Mr. Gurjas Singh Gandhi, Dr. Netraja Mulay
Country	India
Abstract	This paper presents PriceMorphNet, a multimodal deep learning framework for intelligent e-commerce product price prediction using cross-attention fusion of textual and visual embeddings. The proposed architecture integrates 384-dimensional Sentence Transformer text embeddings, 512-dimensional CLIP image embeddings, and engineered categorical features through a cross-modal attention mechanism and deep fusion network. The system is trained on a large-scale dataset containing 75,000 labelled product listings using a Pseudo-Huber loss function in the logarithmic price domain under five-fold cross-validation. Experimental evaluation demonstrates a validation SMAPE of 43.96%, outperforming several uni-modal baseline configurations. The study further validates that combining textual semantics, visual product cues, and structured metadata significantly enhances pricing accuracy. The proposed framework highlights the effectiveness of multimodal representation learning and attention-based fusion for scalable retail price intelligence systems.
Keywords	Multimodal Learning, Product Price Prediction, Cross-Attention Fusion, Deep Learning, E-Commerce Analytics, CLIP, Sentence Transformer, Vision-Language Models, Retail Intelligence, SMAPE
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 8, Issue 3, May-June 2026
Published On	2026-05-29

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research