Document Classification And Automatic Grading
G. L. Sankara Subramaniyan, S. Yajith Vishwa, T. Yogith, Uma K.V, C Deisy · 2022 Second International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT) · 2022
Computing similarity between documents plays a vital role in grade assessment. Methods like word-by-word comparison are available to compute similarity between documents. Automatic assessments have to deal with lots of unstructured data which may be the same in meaning but different in sentences. A new methodology has to be adapted to compute similarity between unstructured documents. This paper provides a faster and efficient way to make embeddings for documents and computes the similarity for grade calculation. The most widely used similarity measure cosine similarity is used for computing the similarity score. To make embeddings for documents, this paper uses two algorithms Word2Vec and Doc2Vec. An aggregate result of these two methods is used for calculating grade scores.