Datasetpublic

Khyatimirani/pcos_question_answer_hindi

PCOS Hindi Lifestyle & Clinical Q&A Dataset Dataset Details Dataset Description This dataset contains patient-facing conversational question–answer pairs in Hindi (Devanagari script) focused on Polycystic Ovary Syndrome (PCOS/PCOD). The dataset is designed to support training and evaluation of healthcare conversational AI systems that provide lifestyle and general clinical guidance for women diagnosed with PCOS. All conversations are structured in a… See the full description on the dataset page: https://huggingface.co/datasets/Khyatimirani/pcos_question_answer_hindi.

sourceHugging Faceapache-2.0updated 7mo agoView on Hugging Face

0likes26downloads

Dataset Card

PCOS Hindi Lifestyle & Clinical Q&A Dataset

Dataset Details

Dataset Description

This dataset contains patient-facing conversational question–answer pairs in Hindi (Devanagari script) focused on Polycystic Ovary Syndrome (PCOS/PCOD).

The dataset is designed to support training and evaluation of healthcare conversational AI systems that provide lifestyle and general clinical guidance for women diagnosed with PCOS.

All conversations are structured in a chat format with user and assistant roles. The content is derived from clinic guidelines, fact sheets, and doctor-reviewed educational material related to PCOS management.

The dataset focuses on:

—Lifestyle management (diet, exercise, sleep, stress)
—General symptom awareness
—Weight management guidance
—Menstrual irregularity information
—Fertility-related concerns (non-diagnostic)
—General medical guidance provided in educational materials

The dataset does not contain diagnostic instructions, personalized prescriptions, or emergency medical advice.

—Curated by: Khyati Mirani
—Language(s) (NLP): Hindi (Devanagari script)
—License: Apache-2.0

Uses

Direct Use

This dataset is intended for:

—Fine-tuning conversational AI models in Hindi
—Post-training alignment of healthcare assistants
—Evaluating medical Q&A generation quality
—Building PCOS-focused digital health assistants
—Research in multilingual medical NLP

Out-of-Scope Use

This dataset is not intended for:

—Direct clinical diagnosis
—Emergency medical decision-making
—Replacing consultation with qualified healthcare professionals
—Generating personalized treatment plans
—Legal or insurance-related medical decision support

Dataset Structure

The dataset is formatted as a JSON array.

Each item follows the structure:

json

{
  "messages": [
    {"role": "user", "content": "Patient question in Hindi"},
    {"role": "assistant", "content": "Assistant response in Hindi"}
  ]
}
## Characteristics:

Fully in Hindi (Devanagari script)

No English words in conversational content

Short, clear responses (2–4 sentences)

No hallucinated medical content

Extractive or guideline-grounded responses only

Designed to reduce over-assertive medical claims

## Dataset Creation

### Curation Rationale

PCOS is a highly prevalent endocrine condition, yet reliable Hindi conversational datasets for medical AI training are limited.

This dataset was created to:

Improve access to structured PCOS information in Hindi

Enable development of culturally relevant health AI systems

Reduce misinformation in regional-language medical assistants

Support safe conversational healthcare AI development

### Source Data

The dataset is derived from:

Doctor-written PCOS lifestyle guidelines

Clinic educational materials

FAQ documents

Hindi-translated patient information sheets

Structured medical fact sheets

Only actionable, patient-facing educational content was used.

Promotional content, testimonials, author biographies, and purely theoretical sections were excluded.
#### Data Collection and Processing

Source documents were reviewed section by section.

FAQ blocks were converted into conversational Q&A format.

Doctor recommendations were reframed into patient-style questions.

Tables and schedules were converted into practical responses.

Language consistency ensured: 100% Hindi output.

No new medical advice was introduced.

Numerical values and durations were preserved exactly as stated in source materials.

No synthetic medical claims were added.

#### Who are the source data producers?

The original materials were created by healthcare professionals and fertility/gynecology clinics. The conversational transformation was performed by the dataset curator.

No personal patient data was used.

### Annotations [optional]

<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->

#### Annotation process

#### Who are the annotators?

This dataset does not contain external annotations such as labels, classifications, or sentiment tags.

It is structured conversational data only.

#### Personal and Sensitive Information

The dataset does not contain:

Personal patient identifiers

Names, addresses, or contact details

Individual medical records

Real patient case histories

All content is educational and general in nature.

## Bias, Risks, and Limitations

## Limitations
Based on specific clinic guidelines; may not represent all medical viewpoints.

Not a substitute for personalized medical consultation.

Cultural or regional bias may exist depending on source material.

Lifestyle advice may not apply universally.

Does not include complex comorbidities.

## Risks

Users may interpret general guidance as personalized advice.

Over-reliance on AI-generated health guidance without consulting doctors.

### Recommendations

Models trained on this dataset should include clear medical disclaimers.

Systems should escalate emergency or severe symptom queries.

Human-in-the-loop validation is recommended for clinical deployment.

Additional datasets should be incorporated for broader medical robustness.

## Citation [optional]

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

@dataset{pcos_hindi_dataset_2026,
  author = {Mirani, Khyati},
  title = {PCOS Hindi Lifestyle & Clinical Q&A Dataset},
  year = {2026},
  license = {Apache-2.0}
}

**APA:**

Mirani, K. (2026). PCOS Hindi Lifestyle & Clinical Q&A Dataset.

## Dataset Card Authors [optional]

Khyati Mirani

## Dataset Card Contact

khyatimirani77@gmail.com