Overview
Although written exams are a more effective method than test exams in terms of seeing the expression of the student with the words used in the evaluation of the competencies related to the comprehension of education, they are not preferred due to the large number of students in our country and the evaluation processes are long, many exams are applied with test methods. For this reason, software and a system that will facilitate the classical (written) exam evaluation processes and provide efficiency is required.
Automatic Exam Evaluation Software with Machine Learning and Natural Language Processing Techniques will only be used for the evaluation of questions consisting of letters, words and sentences, and question answers consisting of numbers, symbols, graphics, drawings and similar characters and content will not be evaluated by the system. In this direction, the system output is used for scoring all Turkish exam questions, which are only composed of letters, have text content, and are within the scope of paragraph type answers. The scoring of the questions is carried out by Machine Learning and Natural Language Processing (NLP) techniques.