Jointly Predicting Job Performance, Personality, Cognitive Ability, Affect, and Well-Being

“…These studies relied on surveys, which can be time-consuming and introduce bias. Various sensors can be leveraged to collect data and automatically predict well-being, e.g., motion and temperature sensors in smart home environments [27,28] or smartphone application usage [29][30][31]. Regarding speech features, to predict well-being in the context of depression, researchers [32][33][34] used speech features obtained from audio and video, such as, interviews and reading tasks.…”

“…Early research on the five-factor model and user behaviour focused only on a limited number of aspects, such as personality and interpersonal relationships, personality and purchase propensity, and personality and career choice, with extraversion and conscientiousness being the most frequently studied personality traits (Forret and Dougherty, 2001;Robles-Granda et al, 2020). Due to the development of online technologies, especially the emergence of social networks, interpersonal communication has changed.…”

“…While this research allowed us to establish how social media data can work in concert with other forms of sensed data, it did not provide insights into the gains it offered over conventional sources. To address this, in subsequent research we developed new methodologies and analytical frameworks that showed that both person-centered analyses of social media data and passive sensor data-augmented social media predictions provided gains over variable-centered predictions that used social media alone (Robles-Granda et al, 2020;Saha et al, 2021). Since precision medicine argues that such predictions be personalized to account for clinical heterogeneity, we thereafter augmented the predictive assessments by first identifying clusters of individuals based on physical sensor data streams, and then building personalized predictive models on these stratified samples, where each stratified cluster represented a different set of lifestyles and behaviors Saha et al, 2021).…”