| 摘要 |
PROBLEM TO BE SOLVED: To analyze multiple pieces of sensor data that are intermittently sent and different in data type, in real time by not using batch processing and regardless of the increase and decrease of the number of pieces of sensor data, and thereby generate precise recommendation information.SOLUTION: Multiple pieces of different types of sensor data (containing operation data of a terminal) are collected to one time series data and are analyzed as mixed time series without separating the pieces of data on the way into every data type. Accordingly, the individual values acquired from each sensor are not separated and analyzed, so that the type and number of various pieces of sensor data are not determined previously and the number of sensors can flexibly be expanded. Thus, the pieces of sensor data are supplemented with each other in time series, and thereby a state of a user can be analyzed in real time in a bird's eye view manner and continuously. A level of abstraction of the data is increased by translating time series data and repeating the processing of generating another new time series data, and thereby the pieces of data of multiple sensor types are processed on the same time series data.SELECTED DRAWING: Figure 1 |
