File size: 2,830 Bytes
e9e7ae3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ce36dfc
e9e7ae3
ce36dfc
 
 
 
 
e9e7ae3
 
 
 
 
 
 
ce36dfc
 
e9e7ae3
2908d8f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7d1b5af
 
 
 
 
 
2908d8f
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
---
dataset_info:
  features:
  - name: tweet_id
    dtype: string
  - name: text
    dtype: string
  - name: label
    dtype:
      class_label:
        names:
          '0': SIN_PROFESION
          '1': CON_PROFESION
  splits:
  - name: train
    num_bytes: 711780
    num_examples: 2786
  - name: validation
    num_bytes: 238488
    num_examples: 999
  - name: test
    num_bytes: 242754
    num_examples: 1001
  download_size: 807660
  dataset_size: 1193022
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---

# Binary Classification Dataset: Profession Detection in Tweets

This dataset is a derived version of the original **PROFNER** task, adapted for binary text classification. The goal is to determine whether a tweet **mentions a profession or not**.

## 🧠 Objective

Each example contains:
- A `tweet_id` (document identifier)
- A `text` field (full tweet content)
- A `label`, which has been normalized into two classes:
  - `CON_PROFESION`: The tweet contains a reference to a profession.
  - `SIN_PROFESION`: The tweet does not contain any profession-related term.

## 📦 Dataset Structure

The dataset is formatted as a `DatasetDict` with three splits:

| Split        | Description                                           |
|--------------|-------------------------------------------------------|
| `train`      | Balanced dataset containing both classes              |
| `validation` | Contains equal distribution of profession/no-profession |
| `test`       | Also balanced for evaluating binary classification    |

Each example follows the structure:

```python
{
  "tweet_id": "1242399976644325376",
  "text": "Nuestros colaboradores y conductores se quedan en casa!",
  "label": "CON_PROFESION"  # or "SIN_PROFESION"
}
```

The `label` column is implemented with Hugging Face `ClassLabel`, which makes it easy to convert between string and integer representation.

## 🔄 Label Mapping

The dataset uses the following class labels:

```python
label_list = ["SIN_PROFESION", "CON_PROFESION"]
label2id = { "SIN_PROFESION": 0, "CON_PROFESION": 1 }
id2label = { 0: "SIN_PROFESION", 1: "CON_PROFESION" }
```

These are automatically applied via Hugging Face `datasets.Features`.

## 📥 How to Load

```python
from datasets import load_dataset

ds = load_dataset("luisgasco/profner_classification_master")
print(ds["train"][0])
# Show features
print(ds["train"].features)
# Ver etiquetas as strings para un ejemplo:
example = ds["train"][5]
print(example["label"])  # IDs
print(ds["train"].features["label"].int2str(example["label"]))
```

## ✍️ Author

Processed and [Luis Gasco](https://huggingface.co/luisgasco) for educational purposes, based on the PROFNER corpus.