Datasets:
metadata
license: mit
task_categories:
- text-classification
language:
- en
tags:
- sentiment-analysis
- text-classification
- multiclass-classification
pretty_name: Sentiment Analysis Preprocessed Dataset including training and testing split
size_categories:
- 10K<n<100K
Brief idea about dataset:
This dataset is designed for a Text Classification to be specific Multi Class Classification, inorder to train a model (Supervised Learning) for Sentiment Analysis.
Also to be able retrain the model on the given feedback over a wrong predicted sentiment this dataset will help to manage those things using Other Features.
Main Features
| text | labels |
|---|---|
| This feature variable has all sort of texts, sentences, tweets, etc. | This target variable contains 3 types of numeric values as sentiments such as 0, 1 and 2. Where 0 means Negative, 1 means Neutral and 2 means Positive. |
Other Features
| preds | feedback | retrain_labels | retrained_preds |
|---|---|---|---|
| In this variable all predictions are going to be stored. | In this variable user can enter either yes or no to indicate whether the prediction is right or wrong. | In this variable user will enter the correct label as a feedback inorder to retrain the model. | In this variable all predictions after feedback loop are going to be stored. |