الفهرس | Only 14 pages are availabe for public view |
Abstract •Study problem: The current study problem statement is how to automatically assess students’ Arabic essays by expert system using data mining? • Study objectives: Automated Arabic essay assessment and scoring is not a trivial problem, since it depends on the capacity of mental abilities like comparisons, understanding and the relations of texts to each other, etc. Therefore, it is an active field in Artificial Intelligence (AI), expert systems, data mining, text mining, and Natural Language Processing (NLP) domains. So, this study defines what assessment is, and presents a background knowledge about these domains. In addition, it gives an overview of current well-known Automated Essays Scoring (AES) systems. The main objective of this study is to develop an expert system to assess students’ Arabic essays called AAES, which achieve the following: 1. Obtaining a method to prepare Arabic text to be processed programmatically. 2. Merging NLP with data mining to score students’ Arabic essays in an expert system. 3. Assigning score of student’s Arabic essay automatically. 4. Assistance teachers when scoring Arabic essays. 5. Saving teachers’ times, and speed up scoring operation. 6. Providing student with his score in a few seconds. 7. Saving education resources. • The research sample: Set of 45 Arabic essays collect from students’ answer in the first grade in ”Computer Teacher Preparation Department” • The study results: The correlation between AAES and the human assessor reached 0.93; that means very good results; whereas the correlation between two human assessors was 0.94. These results indicate that AAES acts as a human assessor to a large extend. • Key words: Automated Arabic Essay Scoring, Data Mining, Natural Language Processing, Latent Semantic Analysis, Expert Systems. |