With the increased diversity among the available data sources, the demand for sufficient storage and processing methods has grown in recent years. Also, extracting the inherent knowledge from the massive as well as heterogeneous information sources becomes an emerging research area. The abundance of the data quantity in the organizations has remarkably raised the gap between the disparate information systems. The advancement of semantic web technologies plays a prominent role in the decision

more »

... ing by providing the contextual information of the data residing in the heterogeneous data sources. Despite, lack of updation with the modifications in the information systems leads to ineffective decision making. To overcome this constraint, the existing researchers have effectively managed the data sources and supported the interoperability among the diverse data sources. Hence, it is significant to integrate the heterogeneous data sources into unified information-rich systems with the help of ontologies. Thus, this work presents heterogeneous data source integration model based on novel semantic ontology. The proposed integration model involves two phases such as novel kernel-based similarity learning and enhanced K-nearest neighbor clustering. By utilizing the novel semantic ontology, it effectively integrates the heterogeneous data sources. Initially, the proposed semantic data integration model obtains the potential information from the data sources such as Header information comprising the IP address, Source details, and destination details. Secondly, it applies the kernel-based similarity learning to compute the similarity between the heterogeneous data sources. Finally, the proposed integration model groups similar data using K-Nearest Neighbor (KNN) based clustering method. Thus, the experimental framework outperforms the existing data based web discovery methods by accomplishing improved performance in terms of the accuracy, success rate and execution time.