Python Based Machine Learning Text Classification

Abstract

Machine learning is a clear indication that technological progress is accelerating right now. All jobs must be done instantly and quickly in the digital world, making businessmen more competitive. There is also a pressing demand for machine learning in education. As far as future policies go, this is relevant. Included in this issue is the challenge of evaluating the substance of a student's thesis abstract for accuracy, appropriateness, and topicality. One method for addressing these issues is classification. The appropriateness level of abstract writing can be determined by classifying the labeled training data according to the topic. Of course, the training and test data will be cleaned or pre-processed prior to classification. Pre-processing and classification tasks benefit from the python library's speed and accuracy. In 2019, 126 abstract papers with 8 labels were used, and in 2020, 116 abstract documents with 8 labels were used. The student's thesis abstract follows the topic with an accuracy of 93.85 percent, a precision of 97 percent, and a recall of 95 percent, according to the results of the study.