Technical Document Classifications Using AI - Statistical Methods

Due to the development of the digital content, there is a high demand for automated document classification systems. In this paper, the author proposes an AI-based statistical approach to categorize documents as technical or nontechnical. The system input are files in PDF format and the output is the probability that the document is of one of the categories. The workflow begins with cleaning and tokenization of the text as are well used as feature for performance a extraction classification of user-friendly where Random the interface categorization we Forest system, which and of use and the is thus fields. methods a metrics developed provides such combination that using a as of are Flask useful BERT Logistic used and tool embeddings. Regression include Streamlit. for The and accuracy organizing following Decision and large machine Tree. precision This sets learning To and paper of models assess the simplifies data the results the with are document potential presented application in in a number.