Segmentation of Older Adults in the Acceptance of Social Networking Sites Using Machine Learning.

This study analyzes the most important predictors of acceptance of social network sites in a sample of Chilean elder people (over 60). We employ a novelty procedure to explore this phenomenon. This procedure performs apriori segmentation based on gender and generation. It then applies the deep learning technique to identify the predictors (performance expectancy, effort expectancy, altruism, telepresence, social identity, facilitating conditions, hedonic motivation, perceived physical condition, social norms, habit, and trust) by segments. The predictor variables were taken from the literature on the use of social network sites, and an empirical study was carried out by quota sampling with a sample size of 395 older people. The results show different predictors of social network sites considering all the samples, baby boomer (born between 1947 and 1966) males and females, silent (born between 1927 and 1946) males and females. The high heterogeneity among older people is confirmed; this means that dealing with older adults as a uniform set of users of social network sites is a mistake. This study demonstrates that the four segments behave differently, and many diverse variables influence the acceptance of social network sites.