Custom Taxonomy Text Classification for Enriched Granularity

Taxonomic text classification is a branch of Natural Language Processing that focuses on organizing textual data into a structured schema. Its applications span various domains, including document management, content recommendation systems, and market research. By integrating custom sub-taxonomies into the existing taxonomies, the research enables a more granular categorization tailored to specific use cases. Leveraging diverse topic modeling techniques and a semi-supervised approach, the project enhances flexibility and performance, facilitating a nuanced understanding of textual information and feature extraction. Performance assessment covers both English and Italian languages, broadening the scope of the experiments. Through rigorous experimentation and comparative analysis, this study identifies the strengths and weaknesses of the current model, providing valuable insights for future investigations.