> For the complete documentation index, see [llms.txt](https://docs.datasaur.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.datasaur.ai/assisted-labeling/ml-assisted-labeling/sentiment-analysis.md).

# Sentiment Analysis

{% hint style="info" %}
**Supported labeling types**: Row labeling.
{% endhint %}

Sentiment analysis classifies text as “Positive” or “Negative”. The sentiment analysis model is based on [DistilBERT-base-uncased-finetuned-SST-2](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english), a fine-tuned version of DistilBERT-base-uncased, specifically trained on the SST-2 dataset.

<figure><img src="/files/rR4ygzFLEGDbAeZ6ydaE" alt="Image of ML Assisted with Sentiment Analysis"><figcaption></figcaption></figure>

### Model details

* The model is using a distilled version developed by Hugging Face based on the Text Classification task pipeline.
* Trained on Stanford Sentiment Treebank ([sst2](https://huggingface.co/datasets/stanfordnlp/sst2)) corpora which contains 67,349 movie review excerpts with human-annotated sentiment labels.
* The model achieves a **91.3%** accuracy on the development set.
* The model hosted locally within the Datasaur Intelligence container.

### Usage

* This model is primarily used for sentiment classification and can also be used for topic classification.
* The base model supports masked language modeling and next sentence prediction but it is primarily intended for fine-tuning on downstream tasks.
* To explore additional fine-tuned versions for different tasks, check out the [Hugging Face model hub](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english).