Designing XML tag based Sindhi language corpus

Corpus play a vital role in building the foundation for processing any language recourse. Sindhi being a very old language, yet still is very limited in computational resources. In this paper, Sindhi corpus construction is discussed using XML tagging, to facilitate Sindhi Natural Language Processing and Machine learning features. A two-chain structure is defined for Sindhi language consisting of metadata and Sindhi source Document tags. These tags are further divided to create a complex XML structure with custom tags to identify each Sindhi document extracted from numerous sources.