SandhiKosh: A Benchmark Corpus for Evaluating Sanskrit Sandhi Tools
17243Authors
nan
Abstract
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 As.t.¯adhy¯ay¯ı by P¯an.ini. This has led to the development of a number of 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. Keywords: Sanskrit, Sandhi, Morphophonology 1.
Keywords
Full Content
SandhiKosh: A Benchmark Corpus for Evaluating Sanskrit Sandhi Tools Shubham Bhardwaj, Neelamadhav Gantayat, Nikhil Chaturvedi, Rahul Garg, Sumeet Agarwal Indian Institute of Technology, IBM Research New Delhi, INDIA shubhamiitd.007@gmail.com, neelamadhavg@in.ibm.com, 709nikhil@gmail.com, rahulgarg@cse.iitd.ac.in, sumeet@ee.iitd.ac.in Abstract 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 As.t.¯adhy¯ay¯ı by P¯an.ini. This has led to the development of a number of 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. Keywords: Sanskrit, Sandhi, Morphophonology 1. Introduction On the 11th of December, 2014, the United Nations Gen- eral Assembly unanimously passed a resolution declaring the summer solstice, June 21st, as the International Day of Yoga (UN, 2014). The resolution recognizes that “yoga provides a holistic approach to health and well-being” and that “a wider dissemination of information about the ben- efits of practicing yoga would be beneficial for the health of the world population”. The popularity of Yoga has been growing steadily all over the world primarily due to its po- tential health and healing benefits in ailments such as de- pression (Uebelacker et al., 2010), cardiovascular disorders (Raub, 2002) and other chronic diseases (Yang, 2007). With the growing worldwide popularity of this ancient dis- cipline, there is also a growing interest to understand and practice this discipline in its pure and unadulterated form. All of the classic texts of this discipline such as Gherand Samhita (Saraswati, 2013), Hath-Yoga-Pradiptika (Singh, 1915), Yoga-Vashistha (Mitra, 1891), Patanjali-Yoga-Sutra (Prasada, 1998) and Bhagavad-g¯ıt¯a (Swarupananda, 2016) were composed in the Sanskrit language. While transla- tions of some of these texts are available in other languages, for a serious Yoga practitioner, reading and understand- ing these texts directly in their native language has a lot of value. Unfortunately, Sanskrit is now spoken by only a small number of people. These classical texts, although available, remain inaccessible to most of the world. Due to this and some more reasons, there is growing in- terest in learning the Sanskrit language (HT, 2007; Ghosh, 2015). For example, the St. James Junior school in Lon- don has introduced Sanskrit language in the junior school because the “knowledge of grammar ultimately gives the pupils a greater clarity and accuracy in thinking, reading and speaking” (SJJS, 2017). Fortunately, the Sanskrit lan- guage has undergone very little modification and by learn- ing this language, it is possible to read and understand most of the classical texts in Sanskrit, including …
Content truncated to 500 words
Labeled Text
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 As.t.¯adhy¯ay¯ı by P¯an.ini. This has led to the development of a number of 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. Keywords: Sanskrit, Sandhi, Morphophonology 1.
Publication Details
- Paper ID:
- 17243
- Year:
- 2018
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan