Automatically Assessing Free Texts

摘要

Evaluation of the content of free texts is a challenging task for humans. Automation of this process is largely useful in order to reduce human related errors. We consider one instance of the "free texts" assessment problems; automatic essay grading where the task is to grade student written essays automatically given course materials and a set of human-graded essays as training data. We use a Latent Semantic Analysis (LSA) -based methodology to accomplish this task. We experiment on a dataset obtained from an occupational therapy course and report the results. We also discuss our findings, analyze different problem areas and explain the potential solutions.