Scalable Daily Human Behavioral Pattern Mining from Multivariate Temporal Data

Abstract

This work introduces a set of scalable algorithms to identify patterns of human daily behaviors. These patterns are extracted from multivariate temporal data that have been collected from smartphones. We have exploited sensors that are available on these devices, and have identified frequent behavioral patterns with a temporal granularity, which has been inspired by the way individuals segment time into events. These patterns are helpful to both end-users and third parties who provide services based on this information. We have demonstrated our approach on two real-world datasets and showed that our pattern identification algorithms are scalable. This scalability makes analysis on resource constrained and small devices such as smartwatches feasible. Traditional data analysis systems are usually operated in a remote system outside the device. This is largely due to the lack of scalability originating from software and hardware restrictions of mobile/wearable devices. By analyzing the data on the device, the user has the control over the data, i.e. privacy, and the network costs will also be removed.

A natural language query interface for searching personal information on smartwatches

“…Moreover, we will develop a retrieval component that can search and show behavioral patterns [20] of the smartwatch to the user. This work focused only on query and not information retrieval.…”

The Motif-Based Approach to the Analysis of the Employee Trajectories within Organization

“…⟨timestamp, employee ID, controlled zone ID, additional data⟩ (1) Additional fields may describe status (entrance permitted or denied), employee access level, and so forth. The logs could be complemented by the control zone plan, including description of the rooms located within a controlled zone, employees' position within organization hierarchy, and the location of the employees' work place within the controlled zones.…”

“…It is necessary to monitor functioning of the technological equipment and people implementing its maintenance. The analysis of the movement of the employees of critical infrastructure and hazardous industries assists in monitoring observance of the safety and access control policies and is useful in detection of insider threat [1,2]. However, existing access control systems and employee activity monitoring systems are mostly targeted for employee productivity assessment by calculating working hours, estimating dynamics of employee lateness, and evaluating user activity on work place [3][4][5].…”

Detecting Transitions in Manual Tasks from Wearables: An Unsupervised Labeling Approach

“…In motif discovery, which has its origins in DNA sequencing research [44,45], time series data are broken down into a set of reoccurring instances, i.e., motifs, by symbolizing the series and then searching for sets of connected symbols. These motifs, sometimes also called frequent patterns or sequences [46,47], can then be applied in several kinds of classification methods. Minnen et al [48] uses motif discovery in a three-step system that first searches for a set of seed motifs, refines these and then trains an HMM for each of the motifs.…”