ScanWise: Concentration-Aware Machine Learning Framework for Cosmetic Product Toxicity Assessment

Assistant Professor Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Associate Professor Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Assistant Professor Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India
Department of Information Science and Engineering Acharya Institute of Technology, Bengaluru, India

This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

The swift expansion of the cosmetics industry gave birth to consumer anxiety about ingredient safety and long-term health effects. The existing cosmetic evaluation tools primarily use fixed ingredient risk databases and simple scoring systems. Concentration often reflects the order, while individual usage pattern is not taken into account by these systems. A dynamic concentration-aware hybrid framework for predicting toxicity of cosmetic products is proposed in this paper. An ingredient toxicity classifier that is based on machine learning is combined with a concentration weighting model. The classifier uses text vectorization via TF-IDF with chosen numeric features to predict toxicity. An exponential decay function is applied to the ingredients to give special importance to the ingredients that appear first in the ingredient list. This is a usage sensitive aggregation which modifies the overall risk of the product as per frequency and intensity. The structure includes a personalized wellness module that checks the compatibility of ingredients with user-specific allergies and skin concerns. Additionally, one component that measures efficacy assesses whether active is recognised. The outcome obtained through analysis of several designs over a curated dataset of cosmetic products demonstrates that concentration aware scoring differentiates risk better than flat-weight scoring methods. All in all, this framework can be used to analyse the safety of cosmetics. It is adaptable and customized.