Abstract
A new methodology called “sensor-based knowledge discovery”, which utilizes wearable sensors and statistical analysis, is proposed and evaluated. This methodology facilitates identifying new knowledge that can improve business outcome. It utilizes wearable sensors to unobtrusively capture people’s location, motion, and social interaction with others. The captured data is converted into multi-dimensional situational variables and then statistically analyzed to deliver a “rule set,” which forms the basis of new knowledge related to business outcome. The methodology was evaluated through a case study at a retail store. A hypothetical rule, that is, a particular area (a so-called “hot spot”) in the store where employee’s presence correlates with average sales per customer, was identified. Based on the identified rule, a measure to concentrate employees in that area was initiated. Consequently, increasing employees’ presence (“staying time”) in the hot spot by 70% increased average sales per customer by 15%. This result demonstrates the effectiveness of the methodology; namely, the new sensor-based knowledge discovery can improve actual business performance.
Recommended Citation
Moriwaki, Norihiko; Hayakawa, Miki; Ohkubo, Norio; Yano, Kazuo; and Senoo, Dai, "Sensor-based Knowledge Discovery from a Large Quantity of Situational Variables" (2013). PACIS 2013 Proceedings. 257.
https://aisel.aisnet.org/pacis2013/257