Tleubayeva commited on
Commit
b772eb1
·
verified ·
1 Parent(s): 882675b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +110 -3
README.md CHANGED
@@ -1,3 +1,110 @@
1
- ---
2
- license: cc-by-4.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ language:
4
+ - kk
5
+ tags:
6
+ - duplicate
7
+ - text-similarity
8
+ - plagiarism-detection
9
+ - kazakh
10
+ - NLP
11
+ size_categories:
12
+ - 10K<n<100K
13
+ ---
14
+ # Dataset Card for KazakhTextDuplicates
15
+
16
+ ## Dataset Details
17
+
18
+ ### Dataset Description
19
+
20
+ The **KazakhTextDuplicates** dataset is a collection of Kazakh-language texts containing duplicates with different levels of modification. The dataset includes **exact duplicates**, **contextual duplicates**, and **partial duplicates**, making it valuable for research in **text similarity, duplicate detection, information retrieval, and plagiarism detection**.
21
+
22
+ - **Developed by:** Arailym Tleubayeva
23
+ - **Language(s) (NLP):** Kazakh (`kk`)
24
+ - **License:** CC-BY-4.0
25
+
26
+
27
+ ## Uses
28
+
29
+ ### Direct Use
30
+
31
+ This dataset can be used for:
32
+ - **Duplicate text detection**: Identifying identical or near-identical texts in large collections.
33
+ - **Plagiarism detection**: Detecting reworded or slightly modified texts.
34
+ - **Information Retrieval (IR)**: Improving search engines by training models to recognize near-duplicate documents.
35
+ - **Semantic Text Similarity (STS)**: Training NLP models to recognize different levels of text similarity.
36
+ - **Data Augmentation**: Generating paraphrased datasets for training models on low-resource Kazakh text data.
37
+
38
+ ### Out-of-Scope Use
39
+
40
+ - Not suitable for **general language modeling** without additional preprocessing.
41
+ - Not designed for **machine translation** tasks.
42
+
43
+ ## Dataset Structure
44
+
45
+ The dataset consists of **Kazakh-language texts** with **three types of duplicate labels**:
46
+
47
+ - **Exact duplicate**: The texts are identical.
48
+ - **Contextual duplicate**: The texts have been slightly modified but retain the same meaning.
49
+ - **Partial duplicate**: Only a portion of the text is similar.
50
+
51
+ ### Columns:
52
+
53
+ | Column Name | Description |
54
+ |------------------|-------------|
55
+ | `id` | Unique document ID |
56
+ | `content` | Original Kazakh text |
57
+ | `category` | Text category (e.g., article, news, academic, etc.) |
58
+ | `language` | Language identifier (`kk` for Kazakh) |
59
+ | `type_duplicate` | Type of duplicate (`exact`, `contextual`, `partial`) |
60
+ | `modified_content` | Modified version of the original text |
61
+
62
+ ## Dataset Creation
63
+
64
+ ### Curation Rationale
65
+
66
+ This dataset was created to address the **lack of publicly available Kazakh-language datasets** for duplicate detection and text similarity tasks. It is particularly useful for NLP applications involving **Kazakh text retrieval, plagiarism detection, and duplicate content filtering**.
67
+
68
+ ### Source Data
69
+
70
+ The dataset consists of **curated text corpora** from:
71
+ - **Public Kazakh-language sources** (news articles, academic texts, Wikipedia, etc.).
72
+ - **Synthetic modifications** created using NLP-based transformations.
73
+
74
+ #### Data Collection and Processing
75
+
76
+ The dataset was processed using:
77
+ - **Text normalization** (lowercasing, punctuation removal).
78
+ - **Stopword removal** (Kazakh stopwords).
79
+ - **Duplication type classification** (exact/contextual/partial).
80
+ - **Data augmentation** for modified versions.
81
+
82
+ #### Who are the source data producers?
83
+
84
+ - Kazakh-language content producers (media, academia).
85
+ - Synthetic text generation tools.
86
+
87
+ ### Annotations [optional]
88
+
89
+ #### Annotation process
90
+
91
+ The dataset was annotated using **automatic similarity detection models** and **manual verification**.
92
+
93
+ #### Who are the annotators?
94
+
95
+ The dataset was reviewed by **Kazakh-speaking NLP researchers** and **linguists**.
96
+
97
+ #### Personal and Sensitive Information
98
+
99
+ The dataset does **not** contain personal or sensitive data. Any potential **identifiable information was removed** during preprocessing.
100
+
101
+ ## Bias, Risks, and Limitations
102
+
103
+ - **Bias in duplicate detection**: The dataset might favor certain duplicate structures over others.
104
+ - **Linguistic variation**: Limited to standard Kazakh; regional dialects may not be fully represented.
105
+ - **Machine-generated paraphrases**: Some `contextual duplicates` were generated using NLP models, which may introduce biases.
106
+
107
+ ### Recommendations
108
+
109
+ Users should be aware of **potential biases** and perform **additional validation** when applying the dataset to real-world tasks.
110
+