A Comprehensive Survey on Methods Used for Automatic Assessment of Subjective Answers

The paper explores various techniques for identifying multilingual text similarities and basic image similarities, covering a wide array of methods and opportunities for comparing and evaluating textual and diagrammatic content. By delving into these categories and studying the pros and cons of various methods, the survey equips researchers, practitioners, and enthusiasts in the field of NLP with a deeper understanding of text similarity analysis.The paper delves into tokenization, vectorization, and similarity metrics, among other approaches. It also discusses modelling techniques, such as LSA and LDA, text alignment algorithms, cosine similarity, BERT, and optimization techniques. The survey categorizes these techniques into corpus-based, knowledge-based, semantic, term-based, and character-based similarities. The survey aims to provide a comprehensive overview of the methods available for checking multilingual text similarity identification and its significance in contemporary text analysis. Index Terms—Similarity Techniques, Machine learning, Natural Language Processing.