Article: M-DRAMA: a multimodal-driven framework for classical drama short video promotion Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.42 pp.18 - 34 Abstract: Facing declining youth engagement in traditional theatre (under 30% attendance), this study addresses the paradox of surging opera-related short video consumption by proposing a multimodal-driven framework for targeted classical drama promotion. We introduce M-DRAMA, an integrated model leveraging three technical innovations: A drama knowledge graph (DKG) with hyperbolic embedding to structure cultural metadata; a cross-modal alignment (CMA) module enforcing frame-level synchronisation of lyrics, movements, and music via matrix constraints, reducing semantic deviation to < 0.2 s. A spatiotemporal interest decoupling network capturing ephemeral youth preferences through gated LSTM-TCN fusion. Validated on the CDS-1K dataset, M-DRAMA achieves NDCG of 0.341 and elevates cultural diffusion index (CDI) by 40%. The framework increases youth user penetration to 37.5%, demonstrating efficacy in minimising cultural discount while balancing algorithmic reach and heritage preservation. Inderscience Publishers - linking academia, business and industry through research

Title: M-DRAMA: a multimodal-driven framework for classical drama short video promotion

Authors: Jun Su

Addresses: College of Humanities and Arts, Xi'an International University, Xi'an, 710000, China

Abstract: Facing declining youth engagement in traditional theatre (under 30% attendance), this study addresses the paradox of surging opera-related short video consumption by proposing a multimodal-driven framework for targeted classical drama promotion. We introduce M-DRAMA, an integrated model leveraging three technical innovations: A drama knowledge graph (DKG) with hyperbolic embedding to structure cultural metadata; a cross-modal alignment (CMA) module enforcing frame-level synchronisation of lyrics, movements, and music via matrix constraints, reducing semantic deviation to < 0.2 s. A spatiotemporal interest decoupling network capturing ephemeral youth preferences through gated LSTM-TCN fusion. Validated on the CDS-1K dataset, M-DRAMA achieves NDCG of 0.341 and elevates cultural diffusion index (CDI) by 40%. The framework increases youth user penetration to 37.5%, demonstrating efficacy in minimising cultural discount while balancing algorithmic reach and heritage preservation.

Keywords: interactive digital media; media convergence; dissemination path optimisation; reinforcement learning; information entropy.

DOI: 10.1504/IJICT.2025.150146

International Journal of Information and Communication Technology, 2025 Vol.26 No.42, pp.18 - 34

Received: 17 Jun 2025
Accepted: 08 Jul 2025
Published online: 01 Dec 2025 *

Title: M-DRAMA: a multimodal-driven framework for classical drama short video promotion

Keep up-to-date