Putting human behavior predictability in context

Abstract

Various studies have investigated the predictability of different aspects of human behavior such as mobility patterns, social interactions, and shopping and online behaviors. However, the existing researches have been often limited to a single or to the combination of few behavioral dimensions, and they have adopted the perspective of an outside observer who is unaware of the motivations behind the specific behaviors or activities of a given individual. The key assumption of this work is that human behavior is deliberated based on an individual’s own perception of the situation that s/he is in, and that therefore it should also be studied under the same perspective. Taking inspiration from works in ubiquitous and context-aware computing, we investigate the role played by four contextual dimensions (or modalities), namely time, location, activity being carried out, and social ties, on the predictability of individuals’ behaviors, using a month of collected mobile phone sensor readings and self-reported annotations about these contextual modalities from more than two hundred study participants. Our analysis shows that any target modality (e.g. location) becomes substantially more predictable when information about the other modalities (time, activity, social ties) is made available. Multi-modality turns out to be in some sense fundamental, as some values (e.g. specific activities like “shopping”) are nearly impossible to guess correctly unless the other modalities are known. Subjectivity also has a substantial impact on predictability. A location recognition experiment suggests that subjective location annotations convey more information about activity and social ties than objective information derived from GPS measurements. We conclude the paper by analyzing how the identified contextual modalities allow to compute the diversity of personal behavior, where we show that individuals are more easily identified by rarer, rather than frequent, context annotations. These results offer support in favor of developing innovative computational models of human behaviors enriched by a characterization of the context of a given behavior.

Using mobile money data and call detail records to explore the risks of urban migration in Tanzania

“…Several studies have used mobile phone data to study mobility and migration patterns in developing countries specifically [ 4 , 55 – 57 ]. Yet despite multi-modal data from different contexts improving prediction accuracy [ 58 ], no research, to the best of our knowledge, has attempted to use this data combined with mobile money transactions to assess the characteristics and potential social and economic consequences of urban migration to the migrant themselves; nor examined the factors that implicate the deprivation level of where people migrate to within urban settings. While academic research on mobile money and migration has previously focused on the rural communities left behind, one prior work [ 59 ] has shown that social networks in a destination location can strongly impact the success of a migration.…”

Lifelong Personal Context Recognition

“…This immediately suggests an ML approach consisting of two steps: collecting examples of sensor recordings annotated with context information and then using this data to learn a map between the two that generalizes to unseen situa-tions. A recent investigation of learning techniques for PCR carried out on the SmartU-nitn2 data shows that indeed automated recognition can be achieved with some degree of success [14,15].…”

“…Figuring out what predictors and architectures are best suited for this task will likely involve borrowing ideas from activity recognition and related areas, while mixing in strategies for dealing with specific aspects of PCR, such as incrementality and support for interaction, which are critical for lifelong alignment (as discussed below). Another important element is that the personal context is inherently structured [15], in the sense that its various aspects are correlated -e.g., a person's activity is strongly influenced by the location that she is in -and constrained by the structure of the context knowledge graph. This hints at the need for developing or repurposing structure-aware predictors.…”

A Context Model for Personal Data Streams

“…1 The app used for the data collection is called iLog [18,17]. The SU data set has been used in a large number of case studies, see, e.g., [19,20]. SU has been collected from one hundred and fifty-eight university students over a period of four weeks.…”