A Talent Evaluation System for Cross-border E-commerce Based on Multimodal Data Fusion
Year: 2026
Type: article
Abstract: The rapid expansion of cross-border e-commerce (CBEC) has highlighted the importance of developing quality human resources that can handle cross-border trade, cultural diversity, and online platforms. Traditional assessment methods focusing on resumes and qualitative assessments are not effective enough to capture a wide range of competencies in international digital trade. Given these issues, the research proposes a multimodal talent assessment system for CBEC that uses textual, audio and video information to provide a comprehensive and fair evaluation. The textual information was lemmatized and processed with Word2Vec, the audio information was filtered and transformed with Short-Time Fourier Transform (STFT), and the video information were extracted using Convolutional Neural Networks (CNN) and did an identification of behavioral cues. For encoding intricate inter-modality interactions, all modalities’ features were combined under the Tensor Fusion Network (TFN) to create a discriminative joint representation. Talent assessment and decision-level classification were done using an Intelligent Fish School Search-driven Attention-fused Gated Recurrent Unit (IFSS-Att-GRU), with attention mechanisms pointing out essential features, and IFSS adjusting model parameters for well-balanced evaluation along skill dimensions. The IFSS-Att-GRU system, implemented in Python, assesses candidates’ professional competence, cross-cultural adaptability, communication effectiveness, and behavioral intelligence in the context of CBEC. The proposed IFSS-Att-GRU + TFN model achieved 0.9632 accuracy, an R² score of 68.45%, an MSE of 0.0056, and an RMSE of 0.0748, outperforming the baseline models by approximately 3–5% in prediction accuracy. Overall, the proposed IFSS-Att-GRU + TFN model contributes an accurate, robust, and scalable multimodal talent evaluation approach that supports objective recruitment and sustainable workforce development in cross-border e-commerce.
Authors: Jingjing Lv, Nan Wang
Institution: Guangdong Ocean University
Language: English
Topic: AI and HR Technologies
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $2,490
OpenAlex ID: W7221587696
FWCI:
Cites:
Cited by:
Related to: