An Enhanced Semantic Retrieval System of Trademarks using Machine Learning

The objectives of paper are to enhance the conceptual based trademark system. Trademarks are essential and important property of a business. A unique trademark allows a company to make status in the market which products or goods it put up for sale. Infringement causes when two trademarks are similar, because of infringement trademarks need security. The conceptual similarity among trademarks arises when more than one trademark evokes the same or similar content. To retrieve this semantic content, there is need of semantic retrieval system of trademarks. Thus, this paper represents an enhancement in semantic retrieval system to find similarity between trademarks using thesaurus of Microsoft word application and feature extraction technique i.e., Principle Component Analysis (PCA) and classification is performed using machine learning algorithms i.e., Artificial Neural Network (ANN), Support Vector Machine (SVM). This system is validated using real 75 infringement cases of trademarks those are conceptual based. The performance is measured using accuracy, precision and recall.