Identification of relations from IndoWordNet for Indian languages using Support Vector Machine

Identification and classification of relations between synsets in a low resource language is a challenging and difficult task, which requires intensive Natural Language Processing (NLP). This paper presents Support Vector Machine (SVM) based approach for learning, classifying and automatically predicting relationships between Hindi Synsets. The average accuracy obtained using SVM is 71.87%, which can be further improved through introduction of language based knowledge. The system performance has been validated using the performance measures namely Precision, Recall and F-score.