Spaces:
Runtime error
Runtime error
Update Space (evaluate main: 05209ece)
Browse files
README.md
CHANGED
@@ -10,6 +10,23 @@ pinned: false
|
|
10 |
tags:
|
11 |
- evaluate
|
12 |
- metric
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
---
|
14 |
|
15 |
# Metric Card for METEOR
|
|
|
10 |
tags:
|
11 |
- evaluate
|
12 |
- metric
|
13 |
+
description: >-
|
14 |
+
METEOR, an automatic metric for machine translation evaluation
|
15 |
+
that is based on a generalized concept of unigram matching between the
|
16 |
+
machine-produced translation and human-produced reference translations.
|
17 |
+
Unigrams can be matched based on their surface forms, stemmed forms,
|
18 |
+
and meanings; furthermore, METEOR can be easily extended to include more
|
19 |
+
advanced matching strategies. Once all generalized unigram matches
|
20 |
+
between the two strings have been found, METEOR computes a score for
|
21 |
+
this matching using a combination of unigram-precision, unigram-recall, and
|
22 |
+
a measure of fragmentation that is designed to directly capture how
|
23 |
+
well-ordered the matched words in the machine translation are in relation
|
24 |
+
to the reference.
|
25 |
+
|
26 |
+
METEOR gets an R correlation value of 0.347 with human evaluation on the Arabic
|
27 |
+
data and 0.331 on the Chinese data. This is shown to be an improvement on
|
28 |
+
using simply unigram-precision, unigram-recall and their harmonic F1
|
29 |
+
combination.
|
30 |
---
|
31 |
|
32 |
# Metric Card for METEOR
|