Multimedia Classification Using DT and SVM-based Multimodal Classifiers: A Metadata-Based Approach

Shwetha, Dr Ashokkumar TA, Jeevitha S

Abstract: The rapid growth of web-based multimedia repositories has created significant challenges in the efficient organization, indexing, and retrieval of digital content. Traditional content-based classification methods often require high computational resources, making metadata-driven approaches an attractive alternative. This study proposes a metadata-based multimedia classification framework that employs two supervised machine learning algorithms, Decision Tree (DT) and Support Vector Machine (SVM), to categorize multimedia video files. Metadata are extracted using the Media Info tool and subsequently preprocessed to remove redundant and irrelevant attributes, improving data quality and classification reliability. Following feature selection, twenty-two significant metadata attributes are retained for model development. The dataset is partitioned into training and testing subsets using a 60:40 ratio to ensure consistent performance evaluation. Both classifiers are assessed using standard metrics, including classification accuracy and confusion matrix analysis. Experimental results demonstrate that metadata provide sufficient descriptive information for effective multimedia classification while eliminating the need for computationally intensive content analysis. Comparative evaluation reveals distinct differences in the predictive capability and classification efficiency of DT and SVM, enabling a comprehensive assessment of their suitability for metadata-driven multimedia categorization. The proposed framework offers a scalable, computationally efficient, and practical solution for multimedia management, supporting applications such as digital libraries, content management systems, multimedia search engines, and intelligent web-based information retrieval.

Keywords: Multimedia Classification, Metadata Extraction, Machine Learning, Decision Tree, Support Vector Machine, MediaInfo, Feature Selection, Web Multimedia, Content Management, Information Retrieval

How to Cite?: Shwetha, Dr Ashokkumar TA, Jeevitha S, "Multimedia Classification Using DT and SVM-based Multimodal Classifiers: A Metadata-Based Approach", Volume 15 Issue 8, August 2026, International Journal of Science and Research (IJSR), Pages: 754-759, https://www.ijsr.net/getabstract.php?paperid=SR26731163202, DOI: https://dx.doi.org/10.21275/SR26731163202

Download Citation: APA | MLA | BibTeX | EndNote | RefMan