Type-2 fuzzy ontology with Dendritic Neural Network based semantic feature extraction for web content classification

Abstract

Nowadays, Internet technology is developing very quickly, because of which webpages are generated exponentially. Web content categorization is mandatory to explore and search related webpages based on queries of users and becomes a dreary task. Most web content categorization methods ignore the contextual knowledge and semantic features of the web page. Pornographic webpage–filtering system does not deliver perfect extraction of advantageous datasets in unstructured web content. Such mechanisms take no reasoning ability to intellectually filter web content to categorize medical websites in adult content webpages. This study introduces a Type-2 Fuzzy Ontology with Dendritic Neural Network Based Semantic Feature Extraction for Web Content Classification (TFODNN-SFEWCC) method. The presented method mainly focused on the detection of different types of web content and blocking pornographic content. It uses the DNN model for the extraction of useful keywords from web pages and eliminates unwanted ones. In addition, the proposed technique employs type 2 fuzzy ontology for the automated classification of web content into multiple classes. The pigeon swarm optimization algorithm is applied to optimize the performance of the Dendritic Neural Network approach for hyperparameter tuning. The experimental evaluation of the proposed method occurs utilizing a web database, and the outcomes are studied under various aspects. The comprehensive comparison study highlighted the betterment of the proposed technique over other existing approaches.

Keywords

• Semantic features
• Web content classification
• Supervised earning
• Pigeon swarm optimizer
• Fuzzy ontology