holylovenia
commited on
Commit
·
8931d52
1
Parent(s):
9b938b5
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- sentiment-analysis
|
4 |
+
language:
|
5 |
+
- ind
|
6 |
+
---
|
7 |
+
|
8 |
+
# indolem_sentiment
|
9 |
+
|
10 |
+
IndoLEM (Indonesian Language Evaluation Montage) is a comprehensive Indonesian benchmark that comprises of seven tasks for the Indonesian language. This benchmark is categorized into three pillars of NLP tasks: morpho-syntax, semantics, and discourse.
|
11 |
+
|
12 |
+
|
13 |
+
|
14 |
+
This dataset is based on binary classification (positive and negative), with distribution:
|
15 |
+
|
16 |
+
* Train: 3638 sentences
|
17 |
+
|
18 |
+
* Development: 399 sentences
|
19 |
+
|
20 |
+
* Test: 1011 sentences
|
21 |
+
|
22 |
+
|
23 |
+
|
24 |
+
The data is sourced from 1) Twitter [(Koto and Rahmaningtyas, 2017)](https://www.researchgate.net/publication/321757985_InSet_Lexicon_Evaluation_of_a_Word_List_for_Indonesian_Sentiment_Analysis_in_Microblogs)
|
25 |
+
|
26 |
+
and 2) [hotel reviews](https://github.com/annisanurulazhar/absa-playground/).
|
27 |
+
|
28 |
+
|
29 |
+
|
30 |
+
The experiment is based on 5-fold cross validation.
|
31 |
+
|
32 |
+
## Dataset Usage
|
33 |
+
|
34 |
+
Run `pip install nusacrowd` before loading the dataset through HuggingFace's `load_dataset`.
|
35 |
+
|
36 |
+
## Citation
|
37 |
+
|
38 |
+
```
|
39 |
+
@article{DBLP:journals/corr/abs-2011-00677,
|
40 |
+
author = {Fajri Koto and
|
41 |
+
Afshin Rahimi and
|
42 |
+
Jey Han Lau and
|
43 |
+
Timothy Baldwin},
|
44 |
+
title = {IndoLEM and IndoBERT: {A} Benchmark Dataset and Pre-trained Language
|
45 |
+
Model for Indonesian {NLP}},
|
46 |
+
journal = {CoRR},
|
47 |
+
volume = {abs/2011.00677},
|
48 |
+
year = {2020},
|
49 |
+
url = {https://arxiv.org/abs/2011.00677},
|
50 |
+
eprinttype = {arXiv},
|
51 |
+
eprint = {2011.00677},
|
52 |
+
timestamp = {Fri, 06 Nov 2020 15:32:47 +0100},
|
53 |
+
biburl = {https://dblp.org/rec/journals/corr/abs-2011-00677.bib},
|
54 |
+
bibsource = {dblp computer science bibliography, https://dblp.org}
|
55 |
+
}
|
56 |
+
```
|
57 |
+
|
58 |
+
## License
|
59 |
+
|
60 |
+
Creative Commons Attribution Share-Alike 4.0 International
|
61 |
+
|
62 |
+
## Homepage
|
63 |
+
|
64 |
+
[https://indolem.github.io/](https://indolem.github.io/)
|
65 |
+
|
66 |
+
### NusaCatalogue
|
67 |
+
|
68 |
+
For easy indexing and metadata: [https://indonlp.github.io/nusa-catalogue](https://indonlp.github.io/nusa-catalogue)
|