MothMalone commited on
Commit
b4e0971
·
verified ·
1 Parent(s): 893c804

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +39 -4
README.md CHANGED
@@ -14,6 +14,21 @@ size_categories:
14
  - 10K<n<100K
15
  - 100K<n<1M
16
  dataset_info:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  - config_name: banking77
18
  features:
19
  - name: text
@@ -67,6 +82,14 @@ dataset_info:
67
  - name: validation
68
  num_examples: 2033
69
  configs:
 
 
 
 
 
 
 
 
70
  - config_name: banking77
71
  data_files:
72
  - split: train
@@ -100,6 +123,7 @@ configs:
100
  - split: validation
101
  path: MASSIVE/validation-*.parquet
102
  ---
 
103
  # Data Preprocessing AutoML Benchmarks
104
 
105
  This repository contains text classification datasets with known data quality issues for preprocessing research in AutoML.
@@ -113,14 +137,25 @@ from datasets import load_dataset
113
 
114
  # Example for loading the TREC dataset
115
  dataset = load_dataset("MothMalone/data-preprocessing-automl-benchmarks", "trec")
 
 
 
116
  ```
117
 
118
  ## Available Datasets
119
 
120
  Below are the details for each dataset configuration available in this repository.
121
 
 
 
 
 
 
 
 
 
122
  ### banking77
123
- - Description:
124
  - Data Quality Issue: N/A
125
  - Classes: 77
126
  - Training Samples: 7502
@@ -136,7 +171,7 @@ Below are the details for each dataset configuration available in this repositor
136
  - Test Samples: 500
137
 
138
  ### financial_phrasebank
139
- - Description: The key arguments for the low utilization of statistical techniques in
140
  - Data Quality Issue: N/A
141
  - Classes: 3
142
  - Training Samples: 1698
@@ -144,9 +179,9 @@ Below are the details for each dataset configuration available in this repositor
144
  - Test Samples: 0
145
 
146
  ### MASSIVE
147
- - Description:
148
  - Data Quality Issue: N/A
149
  - Classes: 60
150
  - Training Samples: 11514
151
  - Validation Samples: 2033
152
- - Test Samples: 2974
 
14
  - 10K<n<100K
15
  - 100K<n<1M
16
  dataset_info:
17
+ - config_name: twenty_newsgroups
18
+ features:
19
+ - name: text
20
+ dtype: string
21
+ - name: label
22
+ dtype: int64
23
+ - name: label_text
24
+ dtype: string
25
+ splits:
26
+ - name: train
27
+ num_examples: 8586
28
+ - name: test
29
+ num_examples: 7532
30
+ - name: validation
31
+ num_examples: 2828
32
  - config_name: banking77
33
  features:
34
  - name: text
 
82
  - name: validation
83
  num_examples: 2033
84
  configs:
85
+ - config_name: twenty_newsgroups
86
+ data_files:
87
+ - split: train
88
+ path: twenty_newsgroups/train.csv
89
+ - split: test
90
+ path: twenty_newsgroups/test.csv
91
+ - split: validation
92
+ path: twenty_newsgroups/validation.csv
93
  - config_name: banking77
94
  data_files:
95
  - split: train
 
123
  - split: validation
124
  path: MASSIVE/validation-*.parquet
125
  ---
126
+
127
  # Data Preprocessing AutoML Benchmarks
128
 
129
  This repository contains text classification datasets with known data quality issues for preprocessing research in AutoML.
 
137
 
138
  # Example for loading the TREC dataset
139
  dataset = load_dataset("MothMalone/data-preprocessing-automl-benchmarks", "trec")
140
+
141
+ # Example for loading the twenty_newsgroups dataset
142
+ dataset = load_dataset("MothMalone/data-preprocessing-automl-benchmarks", "twenty_newsgroups")
143
  ```
144
 
145
  ## Available Datasets
146
 
147
  Below are the details for each dataset configuration available in this repository.
148
 
149
+ ### twenty_newsgroups
150
+ - Description: The 20 Newsgroups dataset is a collection of approximately 20,000 newsgroup documents, partitioned across 20 different newsgroups.
151
+ - Data Quality Issue: N/A
152
+ - Classes: 20
153
+ - Training Samples: 8586
154
+ - Validation Samples: 2828
155
+ - Test Samples: 7532
156
+
157
  ### banking77
158
+ - Description: Dataset of online banking queries annotated with their corresponding intents.
159
  - Data Quality Issue: N/A
160
  - Classes: 77
161
  - Training Samples: 7502
 
171
  - Test Samples: 500
172
 
173
  ### financial_phrasebank
174
+ - Description: Financial sentiment analysis dataset with phrases from financial news.
175
  - Data Quality Issue: N/A
176
  - Classes: 3
177
  - Training Samples: 1698
 
179
  - Test Samples: 0
180
 
181
  ### MASSIVE
182
+ - Description: Multilingual Amazon Slurp Synthetic Intent and Voice Evaluation dataset.
183
  - Data Quality Issue: N/A
184
  - Classes: 60
185
  - Training Samples: 11514
186
  - Validation Samples: 2033
187
+ - Test Samples: 2974