Ontology Construction from Cross Domain Customer Reviews using Expectation Maximization and Semantic Similarity

Recently, researchers are focusing more attention on cross-domain customer review based ontology construction. It seems very little efforts have been taken to construct cross-domain application ontology. Online reviews help customers to get product information in detail. Bringing the reviews of same product from different e-commerce platform into single platform will enable customers make well informed decision. Customers may purchase online product by comparing reviews as manual comparison of reviews is too difficult. In future very large amount of cross-domain applications would be connected to make smart shopping possible. In this paper, two major e-commerce portals, Amazon and Flipkart's reviews are collected, processed and analyzed using expectation maximization (EM) algorithm and semantic similarity. Next, precision and recall, ontology construction and execution time are calculated using state of art models.