Estimation of significant places visited by mobile-terminal user based on communications log to base stations

摘要

A method is disclosed of estimating significant places visited by a mobile-terminal user for wireless communication via base stations. The method includes: collecting at least one communication log represented with consecutive communication events between the mobile terminal and connected one of the base stations in a coverage area of each mobile terminal, the base stations being identified by unique base-station identifiers (BS IDs), respectively, each communication event including date and time of communication and one of the BS IDs which is indicative of the connected base station; dividing each communication log into consecutive time-windowed segments, using a discrete time window moving in time; and, per each time-windowed segment, estimating at least one significant place visited by the user, based on a probability distribution with which the BS IDs appear in each time-windowed segment.

主权项

1. An apparatus for estimating significant places visited by users carrying respective mobile terminals for wireless communication via a plurality of base stations, the apparatus comprising:
a communication-log collector configured to collect, per each mobile terminal, at least one communication log represented with a plurality of consecutive communication events between each mobile terminal and a connected one of the base stations in a coverage area of each mobile terminal, by receiving the communication events from the connected base station, the plurality of base stations being identified by a plurality of unique base-station identifiers (BS IDs), respectively, each communication event including date and time of communication and one of the BS IDs which is indicative of the connected base station; a time-window divider configured to divide each communication log into a plurality of consecutive time-windowed segments, using a discrete time window moving in time, each time-windowed segment including a sub-set of the plurality of communication events, each time-windowed segment represented by a single frequency vector having a plurality of elements allocated to the plurality of BS IDs, respectively, each element of the frequency vector having a value indicative of a frequency with which a corresponding one of the BS IDs appears in the sub-set of communication events belonging to a corresponding one of the plurality of time-windowed segments; a clusterer configured to cluster a plurality of frequency vectors respectively corresponding to the plurality of time-windowed segments, such that a sub-plurality of frequency vectors are clustered together and incorporated into a same cluster, to generate a plurality of clusters, based on co-occurrence of the appeared BS IDs between the plurality of frequency vectors; a stay determination unit configured to determine, per each cluster, whether each user is staying in a coverage area of at least specific one of the base stations, or moving, based on a probability distribution with which the plurality of BS IDs appear in a sub-set of some of the communication events that belong to each cluster; and a significant-place estimator configured to estimate that the coverage area is each user's significant place in life, if the stay determination unit determines that each user is staying in the coverage area, wherein the stay determination unit is further configured to determine, per each cluster, that each user is staying, if the probability distribution indicates that a smaller number of ones of the plurality of BS IDs, which are located in vicinity of an actual location of the mobile terminal, have appeared with a higher probability than those of other BS IDs, and that each user is moving, if the probability distribution indicates that a larger number of ones of the plurality of BS IDs have appeared with a higher probability than those of other BS IDs, and wherein the apparatus operates in a facility that is provided by a telecommunication company or a carrier to collectively manage the plurality of base stations, and that collects from the connected base stations a time series of sets of location data indicative of the actual locations of the connected base stations, at a coarse level of spatial granularity and at irregular time intervals, on a per-mobile-phone basis.