Analysis of Consistency and Structure of Scholarly Papers Using Natural Language Processing

This study aims to analyze author consistency in journal writing and make a simple conclusion by measuring word similarity based on journal content: Title, Abstract, Introduction, and Conclusion. Therefore, this research proposes techniques from Natural Language Processing to clean words, TF-IDF, to determine the vector and cosine similarity, which is used to calculate the similarity of words that are considered similar. All the proposed algorithms can produce consistent values and simple conclusions that are used as accurate information based on data from the identified journal content. The results of this study show that the consistency value after being analyzed is 0.74 or 74%, and a simple conclusion is obtained, which becomes exciting information from the journal content presented.