Социально-демографические предикторы предпочтений при получении рекомендаций по здоровому образу жизни

In recent decades, when people have become aware of the consequences of their habits and actions being reflected in the environment and wellness condition, different alternatives have been developed to improve this issue. Some of the tools are technology-based. Technology has overtaken most of the daily activities; humans have found substitutions to improve their quality of life and facilitate all kinds of processes. This current paper aims to identify predictors that could help target the individuals interested in receiving wellness-related recommendations and use this information as a baseline to develop a mobile application that has the purpose to provide personalized wellness-related recommendations. This mobile application will use machine learning algorithms to optimize the data and provide the most accurate recommendation. The necessary information to conduct this study was collected through a cross-sectional survey conducted in Monteria, Colombia, preceded by a pilot survey that helped us to better understand the needs of the final survey. It is a quantitative and exploratory analysis, where social, demographic, and wellness preference information were collected. On behalf of generating a significant result, further analysis was conducted; firstly, a traditional statistical analysis such as ordinal logistic regression and binary logistic regression. Lastly, modern machine learning analysis such as decision tree (CART) was conducted, finding that it was the most accurate since it does not require a reference variable to generate a prediction. As the study showed, there are social and demographic characteristics that are predictors of preferences related to wellness, and therefore providing personalized wellness recommendations is possible.