In a previous paper, a semantic metadata enrichment software ecosystem(SMESE) based on a multi- platform metadata model and a hybrid machine learning model have been proposed. This work presents the SMESE V3 version enhanced with interest-based enrichments through text analysis approaches for sentiments/emotions detection and hidden topics discovery. SMESE V3 makes it possible to create and use a semantic master catalogue with enriched metadata that allows interest- based search and discovery. This paper presents the design, implementation and evaluation of a the SMESE V3platform using metadata and data from the web, linked open data, harvesting and concordance rules, and bibliographic record authorities. The SMESE V3 platform includes three distinct engines that:
- Identify and enrich sentiment/emotion metadata hidden within the text or multimedia structure using the proposed a new BM-Semantic Sentiment and Emotion Analysis algorithm. 2. Propose an hybrid machine learning model for metadata enrichment.
- 3. Generate semantic to pics by
- text, and multimedia content analysis using the proposed BM- Scalable Annotation-based Topic Detection algorithm.
The performance of SMESE V3is evaluated using a number of prototype simulations by comparing them to existing enriched metadata technique and classifications. The results show that the enhanced SMESE V3 and related algorithms allow greater performance for purposes of interest-based search.
