Automated Essay Scoring — NLP Feature Grading Model

This simulator visualises how an automated essay-scoring system actually works: each essay becomes a point in a three-dimensional feature space — vocabulary diversity, syntactic complexity and semantic coherence — and a tunable weighted model turns those three numbers into a single predicted score. Adjust the feature weights and watch the point cloud recolour and the decision plane retilt in real time; dial in human-rater noise to see how automated grading agreement with a real teacher degrades as human scoring gets less consistent. Live readouts track the average predicted score, the pass rate against your chosen threshold, the Pearson correlation between the AI score and a simulated human score, and the RMSE between the two — the same validation metrics real EdTech teams use before trusting an automated grader in production.