README.md

metadata

license: cc-by-4.0
task_categories:
  - text-classification
  - text-generation
language:
  - en
tags:
  - chemistry
  - biology
  - medical
size_categories:
  - 1K<n<10K

Dataset Card for Cosmetic Ingredient Explanations Dataset (CIED)

This dataset card provides information about the Cosmetic Ingredient Explanations Dataset (CIED), a collection of English-language descriptions for cosmetic ingredients, designed for use in natural language processing (NLP), machine learning (ML), and cosmetic science research and development.

Dataset Details

Dataset Description

The CIED dataset offers detailed explanations of cosmetic ingredients, including their function, benefits, mechanisms of action (where known), and usage considerations. It aims to be a comprehensive resource for understanding the ingredients commonly found in cosmetic and personal care products. The descriptions are written in a clear and accessible style, avoiding overly technical jargon.

• Curated by: Yavuz Yilmaz
• Language(s) (NLP): en (English)
• License: CC BY 4.0 (Creative Commons Attribution 4.0 International)

Uses

Direct Use

The CIED dataset is intended for a variety of NLP and ML applications, including:

• Developing ingredient analysis tools for cosmetic products.
• Building question answering systems to provide information about ingredients.
• Creating personalized skincare recommendation systems.
• Generating summaries of ingredient descriptions.
• Performing sentiment analysis on ingredient descriptions.
• Enabling information retrieval based on ingredient properties.
• Classifying ingredients into categories based on function or chemical structure.
• Training named entity recognition (NER) models to identify ingredients in text.
• Extracting relationships between ingredients and their effects.
• Facilitating machine translation of ingredient information.

Out-of-Scope Use

The CIED dataset is not intended for:

• Providing medical advice or diagnosis. The information in this dataset is for informational purposes only and should not be considered a substitute for professional medical advice.
• Making claims about the efficacy or safety of specific cosmetic products. The dataset focuses on individual ingredients, not formulated products.

Dataset Structure

The dataset is provided in JSON format. Each entry consists of the following fields:

• ingredient: (string) The name of the cosmetic ingredient (typically INCI name).
• description: (string) A detailed explanation of the ingredient.

Example:

{
  "ingredient": "Sodium Hyaluronate",
  "description": "Sodium hyaluronate is the sodium salt of hyaluronic acid. Like hyaluronic acid, it is a powerful humectant that attracts and holds water, providing hydration and plumping the skin. It is often preferred in formulations due to its smaller molecular size, which may enhance penetration."
}

Dataset Creation

Curation Rationale

The CIED dataset was created to address the need for a publicly available, comprehensive, and well-structured dataset of cosmetic ingredient explanations. This resource is intended to facilitate NLP and ML research and development in the cosmetic science domain, where such data is often fragmented, proprietary, or not easily accessible.

Source Data

Data Collection and Processing

The data was collected from multiple publicly available online resources using various web scraping techniques. These resources include:

1. Websites providing information on cosmetic ingredient safety and science: Ingredient names and descriptions were extracted from dedicated ingredient pages on these websites. Web scraping scripts targeted specific HTML elements containing the relevant information.
2. Online dictionaries of cosmetic ingredients: Ingredient names, descriptions, and (where available) ratings or reviews were collected from these resources. The scripts navigated through the structure of these dictionaries (e.g., alphabetical listings) to extract data from individual ingredient pages.

Bias, Risks, and Limitations

• Ingredient Coverage: The dataset may not be exhaustive and might not include every cosmetic ingredient.
• Description Accuracy: While efforts have been made to ensure accuracy and clarity, the descriptions are based on information available at the time of data collection and may be subject to change as scientific understanding evolves.
• Potential Bias in Source Data: The dataset may reflect biases present in the original source materials. For example, certain sources might emphasize particular ingredients or brands.
• Subjectivity in Descriptions: While aiming for objectivity, some level of subjectivity may be present in the descriptions, particularly regarding the perceived benefits of ingredients.

Recommendations

Users should be aware of the potential biases, risks, and limitations outlined above. The CIED dataset should be used as one source of information among many, and it is recommended to consult multiple sources and, when necessary, seek expert advice from dermatologists or other qualified professionals. The dataset is intended for research and development purposes and should not be used to make definitive claims about product safety or efficacy.