Semantic enrichment of mobile phone data records

Abstract

The pervasiveness of mobile phones creates an unprecedented opportunity for analyzing human dynamics with the help of the data they generate. This enables a novel human-driven approach for service creation in a variety of domains (e.g., healthcare, transportation, etc.) Telecom operators own and manage billions of mobile network events (Call Detailed Records -CDRs) per day: interpreting such a big stream of data needs a deep understanding of the events' context through the available background knowledge. We introduce an ontological and stochastic model (HRBModel) to interpret mobile human behavior using merged mobile network data and the geo-referenced background knowledge (e.g., OpenStreetMap, etc.) The model characterizes locations with human activities that can happen (with a given likelihood) there. This allows us to predicatively compile sets of tasks that people are likely to engage in under certain contextual conditions or to characterize exceptional events detected from anomalies in the CDR. An experimental evaluation of the approach is presented.

Deriving human activity from geo-located data by ontological and statistical reasoning

“…Alternatively or complementary, we could use Google Map or Yahoo Map or regional cadastral map database for collecting the POIs within the each cell area. A relevant technique for POIs extraction in a given region from geographical information system is described in [15,45]. Table 1 introduces a brief description of ontology we used to generate useful information from the aforementioned heterogeneous data sources.…”

“…This section introduces a prediction model, called High Level Representation of Behavior Model (HRBModel) [15]. This model generates a set of human activities with a likelihood from the set of POIs.…”

“…The figure shows the most probable activity in top-k is predicted to the user. The preliminary analysis of this evaluation was introduced in our previous work [15]: in this work, we improved the fuzzy reasoning model and re-analyzed the geographical area with richer content. In building this application, first of all, we needed to build a density-based POI distribution grid for this city: for this we collected 4.6 million POIs extracted from the OSM.…”

“…To this purpose, we enrich CDR with human activities, which are derived from our model through location and time. As we mentioned earlier, our main contribution of this article is to use background knowledge regarding human activities that could happen in the context where mobile network events occur [13,14]. The most common mobile network antennas are GSM 900 and GSM 1800.…”

Individualized Time-Series Segmentation for Mining Mobile Phone User Behavior

“…In their approach, they use morning (beginning at 6:00AM and ending at noon), afternoon (ending at 6:00PM), night (all remaining hours) for time segmentation. In [4] In addition to the above time segments, a number of authors [28,29,30] introduce early morning, late morning, midnight and so on statically. For instance, in [31] Such segmentations are also used in various applications such as managing mobile intelligent interruption management system [33], making app prefetch practical on mobile phones [34], mining frequent co-occurrence patterns on the mobile phones [3], mining mobile user habits [35,36].…”