Premium
Development of predictive regression model for perceived hair breakage in Indian consumers
Author(s) -
Kaushik Vaibhav,
Nihul Pratiksha,
Mhaskar Sudhakar
Publication year - 2019
Publication title -
international journal of cosmetic science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.532
H-Index - 62
eISSN - 1468-2494
pISSN - 0142-5463
DOI - 10.1111/ics.12527
Subject(s) - breakage , smoothness , regression analysis , linear regression , statistics , regression , curvature , mathematics , computer science , materials science , mathematical analysis , composite material , geometry
Objective To predict consumer‐perceived hair breakage based on parameters from three distinct categories 1) hair strand parameters‐like curvature, stiffness and tensile strength indices; 2) hair matrix or bulk parameters‐like smoothness, detangling, frizz & volume and; 3) biological factors like age, hair density. Methods Consumer‐relevant evaluation techniques were employed in a uniquely designed protocol to obtain real‐life data from the consumers’ head without impacting or damaging their hair. Hairs of 50 Indian female subjects in the age group of 20–40 years were characterized using various instrumental techniques for parameters mentioned above, apart from the hair breakage count. Multiple regression analysis was performed over the data collected to arrive at a regression equation connecting the hair breakage observed with the key parameters impacting hair breakage. Validation of the model was performed by collecting additional set of hair characterization data for 18 Indian subjects with same recruitment parameters. Results A second order, non‐linear multi‐regression equation was proposed for consumer‐perceived hair breakage with five predictors. A reasonable correlation (R 2 = 0.76) was observed between predicted and observed consumer hair breakage values for the validation set. Apart from the hair surface lubrication parameters (smoothness and detangling forces), inherent extensional strength parameters and biologically relevant parameter – hair density – were found to influence the consumer hair breakage. The proposed model offers different insights into the interplay of parameter. The impact of the key parameters was documented on the consumer hair breakage and the same was found to fit well with the available knowledge. Conclusions Current work demonstrates the usefulness of regression modelling in understanding complex consumer‐relevant parameters by taking a holistic view of consumer hair breakage as a combination of various parameters measured individually at lab scale. The proposed regression equation serves as a tool for product developers to understand the physical parameters of impact when it comes to consumer‐perceived hair breakage and make required changes to the formulation. The method presented can be used to develop model for subjects from other geographies and eventually a generalized model can be proposed.