Article: Optimising the online marketing effectiveness perception using deep neural network integration with semantic mining Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.49 pp.38 - 58 Abstract: Aiming at the problem of prediction deviation caused by ignoring deep semantic information in online marketing effect perception, this study proposes an innovative framework that deeply integrates neural networks and multi-level semantic mining. Traditional methods mostly rely on shallow interaction features, making it difficult to capture complex intentions in texts. Our model achieves deep understanding and alignment of user preferences and product connotations through collaborative fine-tuning of pre-trained language models and graph neural networks. Experiments on public datasets show that, compared with mainstream baseline models, this framework has increased the area under the receiver operating characteristic curve for click-through rate prediction by 2.1% and the ranking metric normalised discounted cumulative gain @10 by 4.7%. All improvements are statistically significant (p < 0.01). This research provides an effective approach for building a more precise and interpretable intelligent marketing system. Inderscience Publishers - linking academia, business and industry through research

Title: Optimising the online marketing effectiveness perception using deep neural network integration with semantic mining

Authors: Lining Wang; Lili Mu

Addresses: Haidu College Qingdao Agricultural University, Laiyang 2652006, China ' Haidu College Qingdao Agricultural University, Laiyang 2652006, China

Abstract: Aiming at the problem of prediction deviation caused by ignoring deep semantic information in online marketing effect perception, this study proposes an innovative framework that deeply integrates neural networks and multi-level semantic mining. Traditional methods mostly rely on shallow interaction features, making it difficult to capture complex intentions in texts. Our model achieves deep understanding and alignment of user preferences and product connotations through collaborative fine-tuning of pre-trained language models and graph neural networks. Experiments on public datasets show that, compared with mainstream baseline models, this framework has increased the area under the receiver operating characteristic curve for click-through rate prediction by 2.1% and the ranking metric normalised discounted cumulative gain @10 by 4.7%. All improvements are statistically significant (p < 0.01). This research provides an effective approach for building a more precise and interpretable intelligent marketing system.

Keywords: online marketing; deep neural networks; DNNs; semantic mining; effect perception; recommendation systems.

DOI: 10.1504/IJICT.2026.153620

International Journal of Information and Communication Technology, 2026 Vol.27 No.49, pp.38 - 58

Received: 31 Dec 2025
Accepted: 28 Jan 2026
Published online: 18 May 2026 *

Title: Optimising the online marketing effectiveness perception using deep neural network integration with semantic mining

Keep up-to-date