TwitterDuyguAnalizi

Model Card for Model ID

<!-- Bu model Türkçe X (Twiter) metinlerini 3 duygu (Olumlu, olumsuz ve Nötr) üzerinden sınıflamaktadır-->

Model Details

Model Description

-The purpose of this model is to perform sentiment analysis on Turkish tweets. It takes tweet text as input and classifies the sentiment into "Olumlu" (Positive), "Olumsuz" (Negative), or "Nötr" (Neutral).

The structure of the model is as follows:

1. It loads a dataset of Turkish tweets with associated sentiment labels.
2. It maps the text labels ("Olumlu", "Olumsuz", "Nötr") to numerical IDs (2, 1, 0).
3. It loads a pre-trained Turkish BERT-based model (dbmdz/bert-base-turkish-cased) for sequence classification.
4. It tokenizes the tweet text using the loaded tokenizer.
5. The dataset is split into training and evaluation sets.
6. A Hugging Face Trainer is used to fine-tune the BERT model on the sentiment analysis task.
7. The model is trained for 3 epochs using specified training arguments (learning rate, batch size, evaluation strategy).
8. Accuracy is used as the evaluation metric.
9. After training, the model is evaluated and saved.
10. Finally, there is an example demonstrating how to use the trained model to predict the sentiment of a new text.

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

• Developed by: [Ass. Prof Engin KARAMAN]

Uses

Foreseeable users of the model:

1. Social media analysts and researchers studying public opinion on Twitter in Turkish.

2. Businesses monitoring brand perception and customer feedback on Twitter in Turkish.

3. Developers building applications that require sentiment analysis of Turkish text, such as content filtering, recommendation systems, or chatbots.

4. Linguists and NLP practitioners interested in Turkish language processing and sentiment analysis.

   Those affected by the model:

5. Individuals whose tweets are analyzed by the model, as their sentiment is being categorized.

6. Businesses and organizations whose social media presence is being analyzed, as insights derived from the model can influence strategic decisions.

7. Users of applications that incorporate this model, as the model's predictions will influence the application's behavior (e.g., content shown, recommendations).

8. Developers and researchers in the Turkish NLP community, as the model can serve as a baseline or component for further research and development.

   Intended use of the model: The primary intended use of this model is to provide an automated way to classify the sentiment of text written in Turkish, specifically targeting the style and context found in tweets. It is designed to assist in understanding the emotional tone of Turkish social media discussions at scale. It can be used for:

• Large-scale sentiment analysis of Twitter data for research or commercial purposes.

• Monitoring and analyzing public sentiment around specific topics, events, or brands.

• Providing sentiment scores for individual tweets or aggregated collections of tweets.

• As a component in larger NLP pipelines or applications requiring Turkish sentiment analysis.

  The model is NOT intended for:

• Making definitive judgments about individuals' true feelings or intentions outside the context of the provided text.

• Analyzing text in languages other than Turkish.

• Analyzing text that significantly deviates from the style and content typically found in Turkish tweets.

• Use in high-stakes decision-making processes without human oversight and validation.

• Replacing nuanced human understanding of complex emotions or sarcasm, which the model may struggle to capture accurately.

More Information [optional]

[[email protected]]