Homework 2 A
Devanshu Arya (10229)
Word Boundary Segmentation
Following are the results obtained for the Word Segmentation using Unigrams. The hindi corpus used is the corpus used in the previous assignment. Segmentation of word boundaries are done on the test file provided and the tabulated results are compared against the groundtruth file provided.
| Precision | Recall | F-Score |
|---|
| 0.98 | 0.94 | 0.96 |
To further test the code, a random groundtruth was created from blogs and newspapers. After removing the spaces from this file, the code was run and the results are tabulated below.
| Precision | Recall | F-Score |
|---|
| 0.98 | 0.93 | 0.96 |
Files
All Files
Codes
Download all the Codes