Sanskrit is an ancient Indian language. Several important texts which are of interest to people all over the world today were written in Sanskrit. The Sanskrit grammar has a precise and complete specification given in the text Astadhyayi by Panini. This has led to the development of a number of {\em Sanskrit Computational Linguistics} tools for processing and analyzing Sanskrit texts. Unfortunately, there has been no effort to standardize and critically validate these tools. In this paper, we develop a Sanskrit benchmark called SandhiKosh to evaluate the completeness and accuracy of Sanskrit Sandhi tools. We present the results of this benchmark on three most prominent Sanskrit tools and demonstrate that these tools have substantial scope for improvement. This benchmark will be freely available to researchers worldwide and we hope it will help everyone working in this area evaluate and validate their tools.
@InProceedings{BHARDWAJ18.755, author = {Shubham Bhardwaj and Neelamadhav Gantayat and Nikhil Chaturvedi and Rahul Garg and Sumeet Agarwal}, title = "{SandhiKosh: A Benchmark Corpus for Evaluating Sanskrit Sandhi Tools}", booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {May 7-12, 2018}, address = {Miyazaki, Japan}, editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and Hélène Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis and Takenobu Tokunaga}, publisher = {European Language Resources Association (ELRA)}, isbn = {979-10-95546-00-9}, language = {english} }
Summary of the paper