Semantics argument classification using Word/POS in constituent and left argument features

Semantic argument classification is process to classify argument refers to the meaning of the argument. In the process of the semantic argument classification used the baseline features and enhance with First and Last Word/POS (Part of Speech) in Constituent features to improve semantic argument classification. In addition to these features, authors add Left Argument feature that utilizes the previous constituent label. Experiment result using Support Vector Machine (SVM), our addition features (First and Last Word/POS in Constituent features and Left Argument features) increase the effectiveness F1 7.27% compare to baseline features only.