Link Prediction in Knowledge Graphs
My MSc thesis presents a survey and experimental comparison of six knowledge-graph embedding models—TransE, ComplEx-N3, TuckER, SimplE, CrossE and TorusE—for predicting missing relationships. I evaluated the models on FB15K, FB15K-237, WN18 and WN18RR using MRR and Hits@K, examining dataset structure, inverse-relation leakage, generalization and the reproducibility of KGE experiments.