Search

Mit Logfiles werden automatisch erstellte digitale Dateien (Thimm/Nehls, Kapitel 69 in diesem Band) bezeichnet, in denen bestimmte Ereignisse elektronisch aufgezeichnet werden. Im sogenannten Web Usage Mining (Srivastava et al. 2000) sind vor allem Web Server Logs von Interesse und werden daher oft synonym zum Begriff „Logfiles“ verwendet. Diese dokumentieren die Zugriffe auf Webseiten, um darauf basierend Webseitenstatistiken erstellen und auswerten zu konnen.

We propose different behaviour and interaction related indicators of artificial actors (bots) and show how they can be separated from natural users in a virtual dating market. A finite mixture classification model is applied on the different behavioural and interactional information to classify users into bot vs. non-bot-categories. Finally the validity of the classification model and the impact of bots on sociodemographic distributions and scientific analysis is discussed.

Social reporting is based on the idea that the members of a location-based social network observe real-world events and publish reports about their observations. Application scenarios include crisis management, bird watching or even some sorts of mobile games. A major issue in social reporting is the quality of the reports. We propose an approach to the quality problem that is based on the reciprocal confirmation of reports by other reports. This contrasts with approaches that require users to verify reports, that is, to explicitly evaluate their veridicality. We propose to use spatio-termporal proximity as a first criterion for confirmation and social distance as a second one. By combining these two measures we construct a graph containing the reports as nodes connected by confirmation edges that can adopt positive as well as negative values. This graph builds the basis for the computation of confirmation values for individual reports by different aggregation measures. By applying our approach to two use cases, we show the importance of a weighted combination, since the meaningfulness of the constituent measures varies between different contexts.