A New Approach for Authorship Verification Using Information Retrieval Features

Abstract

Authorship Verification is the process of verifying an author by checking whether the document is written by the suspected author or not. The performance of the Authorship Verification mainly depends on the features used for differentiating the writing style in the documents. The researchers extracted various types of stylistic features for author identification based on the writing style of a particular author. In this work, a new approach is proposed by using the information retrieval features for author identification. In this approach, the test document was treated as a query and the training data as a document database. Cosine similarity measure is used to compute the similarity between the test document and training data. The accuracy of author verification depends on the scores of the similarity measures. The proposed approach performance is good when compared with most of the state-of-the-art approaches in Authorship Verification.