Update README.md
Browse files
README.md
CHANGED
|
@@ -18,46 +18,8 @@ Please follow the [ALBERT-Persian](https://github.com/m3hrdadfi/albert-persian)
|
|
| 18 |
It aims to classify text, such as comments, based on their emotional bias. We tested three well-known datasets for this task: `Digikala` user comments, `SnappFood` user comments, and `DeepSentiPers` in two binary-form and multi-form types.
|
| 19 |
|
| 20 |
|
| 21 |
-
### DeepSentiPers
|
| 22 |
-
|
| 23 |
-
which is a balanced and augmented version of SentiPers, contains 12,138 user opinions about digital products labeled with five different classes; two positives (i.e., happy and delighted), two negatives (i.e., furious and angry) and one neutral class. Therefore, this dataset can be utilized for both multi-class and binary classification. In the case of binary classification, the neutral class and its corresponding sentences are removed from the dataset.
|
| 24 |
-
|
| 25 |
-
**Binary:**
|
| 26 |
-
1. Negative (Furious + Angry)
|
| 27 |
-
2. Positive (Happy + Delighted)
|
| 28 |
-
|
| 29 |
-
**Multi**
|
| 30 |
-
1. Furious
|
| 31 |
-
2. Angry
|
| 32 |
-
3. Neutral
|
| 33 |
-
4. Happy
|
| 34 |
-
5. Delighted
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
| Label | # |
|
| 38 |
-
|:---------:|:----:|
|
| 39 |
-
| Furious | 236 |
|
| 40 |
-
| Angry | 1357 |
|
| 41 |
-
| Neutral | 2874 |
|
| 42 |
-
| Happy | 2848 |
|
| 43 |
-
| Delighted | 2516 |
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
**Download**
|
| 47 |
-
You can download the dataset from:
|
| 48 |
-
|
| 49 |
-
- [SentiPers](https://github.com/phosseini/sentipers)
|
| 50 |
-
- [DeepSentiPers](https://github.com/JoyeBright/DeepSentiPers)
|
| 51 |
-
|
| 52 |
-
|
| 53 |
## Results
|
| 54 |
-
|
| 55 |
-
The following table summarizes the F1 score obtained as compared to other models and architectures.
|
| 56 |
-
|
| 57 |
-
| Dataset | ALBERT-fa-base-v2 | ParsBERT-v1 | mBERT | DeepSentiPers |
|
| 58 |
-
|:------------------------:|:-----------------:|:-----------:|:-----:|:-------------:|
|
| 59 |
-
| SentiPers (Multi Class) | 66.12 | 71.11 | - | 69.33 |
|
| 60 |
-
| SentiPers (Binary Class) | 91.09 | 92.13 | - | 91.98 |
|
| 61 |
|
| 62 |
|
| 63 |
### BibTeX entry and citation info
|
|
|
|
| 18 |
It aims to classify text, such as comments, based on their emotional bias. We tested three well-known datasets for this task: `Digikala` user comments, `SnappFood` user comments, and `DeepSentiPers` in two binary-form and multi-form types.
|
| 19 |
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
## Results
|
| 22 |
+
The model obtained an F1 score of 87.56% for a composition of all three datasets.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
|
| 24 |
|
| 25 |
### BibTeX entry and citation info
|