system HF staff commited on
Commit
9bb27f9
1 Parent(s): 0f99374

Update files from the datasets library (from 1.2.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.2.0

Files changed (1) hide show
  1. README.md +120 -14
README.md CHANGED
@@ -1,17 +1,123 @@
1
  ---
2
- language:
3
- - BCP-47: fr
4
- - cardinality: 1lng
5
- task_structure: Te2Cl
6
- purpose: SA
7
- size:
8
- - train: 120k
9
- - val: 20k
10
- - test: 20k
11
- language_producers: fnd_l
12
- annotation: no_a
13
- licenses:
14
- src_dataset: none
 
 
 
 
 
15
  ---
16
 
17
- # Dataset Card for Allociné Dataset
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - found
6
+ languages:
7
+ - fr
8
+ licenses:
9
+ - mit
10
+ multilinguality:
11
+ - monolingual
12
+ size_categories:
13
+ - 100K<n<1M
14
+ source_datasets:
15
+ - original
16
+ task_categories:
17
+ - text-classification
18
+ task_ids:
19
+ - sentiment-classification
20
  ---
21
 
22
+ # Dataset Card for Allociné
23
+
24
+ ## Table of Contents
25
+ - [Tasks Supported](#tasks-supported)
26
+ - [Purpose](#purpose)
27
+ - [Languages](#languages)
28
+ - [People Involved](#who-iswas-involved-in-the-dataset-use-and-creation)
29
+ - [Data Characteristics](#data-characteristics)
30
+ - [Dataset Structure](#dataset-structure)
31
+ - [Known Limitations](#known-limitations)
32
+ - [Licensing information](#licensing-information)
33
+
34
+ ## Tasks supported:
35
+ ### Task categorization / tags
36
+
37
+ Text to binary text classification
38
+
39
+ ## Purpose
40
+
41
+ The Allociné dataset was developed for large-scale sentiment analysis in French.
42
+
43
+ ## Languages
44
+ ### Per language:
45
+
46
+ The BCP-47 code for French is fr. Dialect information is unknown (see Speaker section for further details).
47
+
48
+ ## Who is/was involved in the dataset use and creation?
49
+ ### Who are the dataset curators?
50
+
51
+ The Allociné dataset was collected by Théophile Blard.
52
+
53
+ ### Who are the language producers (who wrote the text / created the base content)?
54
+
55
+ The dataset contains movie reviews collected from [Allociné.fr](https://www.allocine.fr/). The content of each review may include information and opinions about the film's actors, film crew, and plot.
56
+
57
+ ### Who are the annotators?
58
+
59
+ No annotations were included in this dataset.
60
+
61
+ ## Data characteristics
62
+
63
+ The texts are movie reviews written by members of the [Allociné.fr](https://www.allocine.fr/) community for various films. The reviews were written between 2006 and 2020. Further information on the kinds of films included in the dataset has not been documented.
64
+
65
+ ### How was the data collected?
66
+
67
+ The reviews and ratings were collected using a list of [film page urls](https://github.com/TheophileBlard/french-sentiment-analysis-with-bert/blob/master/allocine_dataset/allocine_films_urls.txt) and the [allocine_scraper.py](https://github.com/TheophileBlard/french-sentiment-analysis-with-bert/blob/master/allocine_dataset/allocine_scraper.py) tool. Up to 30 reviews were collected for each film.
68
+
69
+ ### Normalization information
70
+
71
+ The reviews were originally labeled with a rating from 0.5 to 5.0 with a step of 0.5 between each rating. Ratings less than or equal to 2 are labeled as negative and ratings greater than or equal to 4 are labeled as positive. Only reviews with less than 2000 characters are included in the dataset.
72
+
73
+ ### Annotation process
74
+
75
+ No annotations were included in this dataset.
76
+
77
+ ## Dataset Structure
78
+ ### Splits, features, and labels
79
+
80
+ The Allociné dataset has 3 splits: _train_, _validation_, and _test_. The splits contain disjoint sets of movies. The following table contains the number of reviews in each split and the percentage of positive and negative reviews.
81
+ Dataset Split | Number of Instances in Split | Percent Negative Reviews | Percent Positive Reviews
82
+ --------------|------------------------------|--------------------------|-------------------------
83
+ Train | 160,000 | 49.6% | 50.4%
84
+ Validation | 20,000 | 51.0% | 49.0%
85
+ Test | 20,000 | 52.0% | 48.0%
86
+
87
+ Each data instance contains the following features: _review_ and _label_. In the Hugging Face distribution of the dataset, the _label_ has 2 possible values, _0_ and _1_, which correspond to _negative_ and _positive_ respectively.
88
+
89
+ ### Span indices
90
+
91
+ No span indices are included in this dataset.
92
+
93
+ ### Example ID
94
+
95
+ The ID is an integer starting from 0. It has no inherent meaning.
96
+
97
+ ### Free text description for context (e.g. describe difference between title / selftext / body in Reddit data) and example
98
+
99
+ For each ID, there is a string for the review and an integer for the label. See the [Allociné corpus viewer](https://huggingface.co/datasets/viewer/?dataset=allocine) to explore more examples.
100
+
101
+ ID | Review | Label
102
+ ---|--------|-------
103
+ 4 | Premier film de la saga Kozure Okami, "Le Sabre de la vengeance" est un très bon film qui mêle drame et action, et qui, en 40 ans, n'a pas pris une ride. | 1
104
+ 5 | L'amnésie est un thème en or pour susciter le mystère. Encore faut-il être capable de construire un scénario qui se tienne. Celui-ci est boursouflé et accumule incohérences et invraisemblances. Notons aussi la stupidité du titre français, sans lien avec l'histoire. | 0
105
+
106
+
107
+ ### Suggested metrics / models:
108
+
109
+ [tf-allociné](https://huggingface.co/tblard/tf-allocine) achieves 97.44% accuracy on the test set.
110
+
111
+ ## Known Limitations
112
+ ### Known social biases
113
+
114
+ The social biases of this dataset have not yet been investigated.
115
+
116
+ ### Other known limitations
117
+
118
+ The limitations of the Allociné dataset have not yet been investigated, however [Staliūnaitė and Bonfil (2017)](https://www.aclweb.org/anthology/W17-5410.pdf) detail linguistic phenomena that are generally present in sentiment analysis but difficult for models to accurately label, such as negation, adverbial modifiers, and reviewer pragmatics.
119
+
120
+ ## Licensing information
121
+
122
+ The Allociné dataset is licensed under the [MIT License](https://opensource.org/licenses/MIT).
123
+