Date Published: April 24, 2019
Publisher: Public Library of Science
Author(s): Zaoyi Sun, Pei Zhang, Zhiwei Ji, Chuansheng Chen, Qun Wan, Xiuying Qian, Peter Karl Jonason.
Previous studies showed that individuals’ traits could be used to explain the similarity of behavioral patterns across different occasions. Such studies have typically focused on personality traits, and have not been extended to psychological needs. Our study used a large dataset of 1,715,078 anonymous users’ App usage records to examine whether the individual’s needs-based profiles of App usage were consistent across different situations (as indexed by categories of App functions). Results showed a high level of consistency across situations in a user’s choice of Apps based on the needs the Apps could satisfy. These results provide clear evidence in support of cross-category App recommendation systems.
Humans have a set of universal basic needs that play important functions in daily life . As one of the earliest theorists of human needs, Henry Murray in 1938 identified five categories of 17 needs: ambition, materialism, power, affection, and information . In 1954, Abraham Maslow conceptualized five levels of human needs: physiological, safety, social (love and belonging), esteem, and self-actualization needs . In 1967, David McClelland identified three needs (achievement, affiliation, and power) as the main motivators for human behavior . Recently, Edward Deci and Richard Ryan proposed self-determination theory that focuses the needs for competence), autonomy, and psychological relatedness [4, 5]. However the structure of needs is conceptualized, there is a general consensus among researchers that psychological needs are deep-rooted in our evolutionary history but serve as the driving force behind modern human behaviors [6, 7].
Using three sources of data, namely, App’s needs probability distribution from , anonymous users’ App usage records, and the functional categories of the Apps, we found that most users showed a high level of consistency in their needs profiles (or the traits of psychological needs) across different situations (indexed by App functional categories). The results demonstrated that App users’ needs profiles may be considered as stable traits, which can be used to predict user behaviors and to develop App recommendation systems across categories.