Automatic Classification of Intellectual Property Legal Cases based on Support Vector Machine

Abstract

The importance of intellectual property in economic development is growing. Support vector machines have produced cutting-edge outcomes in a variety of applications, including document classification. However, existing study used SVM for the IP classification task but it did not produce as excellent results as alternative learning algorithms like Random Forest and KNN. This is because kernel patent classification differs from traditional classification in many ways. We assess the new methods by classifying the international patent collection of documents using the Gaussian Kernel Support Vector Machine (GKSVM). This study looks at how to recognize specific elements in court decision texts automatically and evaluates how important a role they play. In this paper, we used common classifiers to classify patent documents. The proposed classification method, GKSVM, yields the best results, and the evaluation result shows accuracy for the test set sample.