Analysis of Taste Preferences Influenced by Environmental Factors using Multinomial Logistics Regression
Keywords:
Taste Preferences, Environmental Factors, Chi-square Test, Multinomial Logistics Regression, AIC Value, BIC value, Log-likelihood, DevianceAbstract
Environmental exposure influences taste preferences through repeated dietary experiences and social learning. Individuals raised in any region with strong culinary traditions often develop preferences aligned with locally available ingredients and traditional cooking taste preferable. For example, communities living in hot and humid climates tend to prefer spicy or strongly flavored foods, partly due to their preservative and appetite-stimulating properties. This study intends to study impact of environmental factors on individual taste preferences, investigating how external surroundings and situational contexts shape culinary choices. This study uses Taste Preferences dataset that dives into a unique synthetic dataset that predicts food taste preferences (Sweet, Spicy, Sour, or Salty) based on lifestyle habits, sleep cycles, exercise levels, climate zones, and cultural cuisine exposure. To analyze relationships, a quantitative approach was employed using a Chi-square test to explore the association between taste preferences and environmental factors, followed by multinomial logistic regression to model the probability of specific taste preferences based on varying environmental predictors. The models predict the probability of a person belonging to a specific "Taste Preference" category compared to a reference category which is the historical cuisine. The model will check using Log-Likelihood, deviance, AIC, BIC. Normally, values of log-likelihood close to zero are indicative of better fit to data Lesser deviance indicates a better-fitting model, AIC and BIC serve as information criteria which researchers use to select their models. Overall, this study provides a scientific bridge between behavioural psychology and sensory science. By using multinomial logistic regression, you are moving beyond just saying "environment matters" to quantifying the probability of how specific surroundings change what we want to eat.



