English-Hindi anusaaraka is being supported by Satyam Computer Services Limited.
To know more about anusaaraka, please visit anusaaraka home page at http://ltrc.iiit.net/~anusaaraka.
Any machine translation system consists of several modules. However these modules are not robust. They are prone to errors. When a input in typical machine translation system passes through all these modules, the errors get cascaded, and output thus produced is many times meaningless.
Anusaaraka interface presents the output at all these intermediate layers in a readble format in various rows. The user has complete control over the display of rows.
| on | the | bank | Original English |
| ->para | -- | kinaaraa | After Word Sense Disambiguation |
| --+2 | --- | kinaare_para | After Prep Movement |
Use the Show/Hide button in the bottom frame to control the display of rows.
The colors in the output indicate category of the word(s). Following color scheme has been used.
| noun-phrase |
| verb-phrase |
| adjectival-phrase |
| adverbial-phrase |
| preposition |
| conjunctions |
| interjections |
The categories are marked by the POS taggers and the chunkers. One can see the actual POS tags and the chunker markings by unhiding the corresponding rows in the output.
The numbers in lightgreen strip indicate the Paragraph, Sentence, and Row numbers. For example 10.5.G stands for the Gth row of the 5th sentence in 10th paragraph. One can use paragraph and sentence number information to jump to a particular location using GOTO button in the top frame. Thus if you want to jump to say 5th sentence in the 10th paragraph, just type in 10.5 and click on GOTO button.
Your suggestions for improving the anusaaraka output quality are most welcome.
To record your suggestion, open the suggestion row. Go to the word for which you want to give your suggestion. Type in your suggestions in the suggestion row and to submit, press enter.
Last modified: 27 May 2005 12:14 PM