Navi004/DDCBot-mental-disorders

• —annotations_creators:
• —original
• —language:
• —en
• —license: cc-by-4.0
• —multilinguality: monolingual
• —pretty_name: DDCBot Mental Disorders Dataset (Split)
• —size_categories:
• —100K<n<1M
• —source_datasets:
• —Kanakmi/mental-disorders
• —task_categories:
• —text-classification
• —task_ids:
• —multi-class-classification ---

DDCBot Mental Disorders Dataset (Split Batches)

This dataset is derived from `Kanakmi/mental-disorders`, and has been split into 10 batches of 50,000 records each to facilitate efficient Large Language Model (LLM) fine-tuning for the DDCBot project.

🧠 About DDCBot

DDCBot (Disorder Detection & Care Bot) is an intelligent mental health assistant that uses fine-tuned LLMs to identify patterns in mental health-related textual data. It aims to:

• —Detect early signs of various disorders.
• —Provide classification support for mental health professionals.
• —Understand linguistic features in user expressions that may relate to psychological conditions.

📊 Dataset Structure

The dataset is split into the following subsets:

• —batch_1 to batch_10: 50k training records each from the Kanakmi/mental-disorders dataset
• —validation: Original validation set
• —test: Original test set

Each batch (e.g., batch_1, batch_2, ..., batch_10) contains ~50,000 samples with the following fields:

• —text: The user's expression or social media post.
• —label: An integer representing one of several mental health conditions (e.g., depression, anxiety, PTSD, etc.).

🔢 Label Classes

The dataset includes multiple classes such as: Labels:

• —0:'BPD'
• —1:'bipolar'
• —2:'depression'
• —3:'Anxiety'
• —4:'schizophrenia'
• —5:'mentalillness'

(Label mapping can be extracted from the original dataset.)