Constructed a structured heterogeneous text corpus graph to transform text classification problem into a node classification problem. Created semantic rich features by using Text GCN and topic modeling based approach-LDA which are then fed into a novel classification model.