Automated Grading of PowerPoint Presentations Using Latent Semantic Analysis

Manual grading of students’ work takes a long time and it is stressful. Evaluator may be holistic or analytic, lenient non-lenient, experienced inexperienced; which leads to non-uniformity in the assessment. Therefore, essential do automated students' overcome human inadequacies through uniform assessment also, reduces workload evaluators. A novel automatic PowerPoint presentation skills using Latent Semantic Analysis (LSA) proposed. Program implemented python extract features corresponding text appearance, graphics, footer, hyperlink from presentations. presentations are represented feature vectors Space Singular Value Decomposition (SVD). SVD reveals relationships between The grades for evaluated by finding Cosine similarity with reference k number nearest such used grade Kneighbors classifier find neighbors. Similarity approach give 90.90% 81.81% accuracy, respectively, while predicting