Florian Lemmerich

VIKAMINE: open-source subgroup discovery, pattern mining, and analytics

This paper presents an overview on the VIKAMINE system for subgroup discovery, pattern mining and analytics. As of VIKAMINE version 2, it is implemented as rich-client platform (RCP) application, based on the Eclipse framework. This provides for a highly-configurable environment, and allows modular extensions using plugins. We present the system, briefly discuss exemplary plugins, and provide a sketch of successful applications.

Generic pattern trees for exhaustive exceptional model mining

Exceptional model mining has been proposed as a variant of subgroup discovery especially focusing on complex target concepts. Currently, efficient mining algorithms are limited to heuristic (non exhaustive) methods. In this paper, we propose a novel approach for fast exhaustive exceptional model mining: We introduce the concept of valuation bases as an intermediate condensed data representation, and present the general GP-growth algorithm based on FP-growth. Furthermore, we discuss the scope of the proposed approach by drawing an analogy to data stream mining and provide examples for several different model classes. Runtime experiments show improvements of more than an order of magnitude in comparison to a naive exhaustive depth-first search.

Exploratory pattern mining on social media using geo-references and social tagging information

This paper presents exploratory pattern mining techniques for describing communities of resources (e.g., images) and for characterising locations of interest. We utilise tagging information and collaborative geo-reference annotations for characterising resources locations by a set of descriptive patterns. The methods are embedded into an interactive approach for mining, browsing and visualising a set of patterns. As an exemplary use case, we focus on the social photo sharing application Flickr. Utilising publicly available real-world data from this platform, we provide a structural evaluation of the automatic approach as well as an exemplary case study for demonstrating the effectiveness and validity of the interactive approach.

Discovering and Characterizing Mobility Patterns in Urban Spaces: A Study of Manhattan Taxi Data

Nowadays, human movement in urban spaces can be traced digitally in many cases. It can be observed that movement patterns are not constant, but vary across time and space. In this work, we characterize such spatio-temporal patterns with an innovative combination of two separate approaches that have been utilized for studying human mobility in the past. First, by using non-negative tensor factorization (NTF), we are able to cluster human behavior based on spatio-temporal dimensions. Second, for characterizing these clusters, we propose to use HypTrails, a Bayesian approach for expressing and comparing hypotheses about human trails. To formalize hypotheses, we utilize publicly available Web data (i.e., Foursquare and census data). By studying taxi data in Manhattan, we can discover and characterize human mobility patterns that cannot be identified in a collective analysis. As one example, we find a group of taxi rides that end at locations with a high number of party venues on weekend nights. Our findings argue for a more fine-grained analysis of human mobility in order to make informed decisions for e.g., enhancing urban structures, tailored traffic control and location-based recommender systems.

Photowalking the City: Comparing Hypotheses About Urban Photo Trails on Flickr

Understanding human movement trajectories represents an important problem that has implications for a range of societal challenges such as city planning and evolution, public transport or crime. In this paper, we focus on geo-temporal photo trails from four different cities (Berlin, London, Los Angeles, New York) derived from Flickr that are produced by humans when taking sequences of photos in urban areas. We apply a Bayesian approach called HypTrails to assess different explanations of how the trails are produced. Our results suggest that there are common processes underlying the photo trails observed across the studied cities. Furthermore, information extracted from social media, in the form of concepts and usage statistics from Wikipedia, allows for constructing explanations for human movement trajectories.

What Makes a Link Successful on Wikipedia

While a plethora of hypertext links exist on the Web, only a small amount of them are regularly clicked. Starting from this observation, we set out to study large-scale click data from Wikipedia in order to understand what makes a link successful. We systematically analyze effects of link properties on the popularity of links. By utilizing mixed-effects hurdle models supplemented with descriptive insights, we find evidence of user preference towards links leading to the periphery of the network, towards links leading to semantically similar articles, and towards links in the top and left-side of the screen. We integrate these findings as Bayesian priors into a navigational Markov chain model and by doing so successfully improve the model fits. We further adapt and improve the well-known classic PageRank algorithm that assumes random navigation by accounting for observed navigational preferences of users in a weighted variation. This work facilitates understanding navigational click behavior and thus can contribute to improving link structures and algorithms utilizing these structures.

Visual Positions of Links and Clicks on Wikipedia

In this work, we study the visual position of links and their clicks on Wikipedia, particularly where links are visually located, at which screen positions users click on links, and which areas on the screen exhibit more or less clicks per links. For that purpose, we introduce a novel dataset containing the on-screen coordinate position for all links between pages in the English Wikipedia and additionally resort to navigation logs of Wikipedia users. Using this data, we can observe a preference of certain link and click locations on Wikipedia including first evidence of positional click bias. For example, our results suggest that users have a tendency to prefer to click on the left side of the screen which exceeds what one would expect from the presence of links on pages. We believe that presented data and research can be useful for optimizing the process of link creation and link consumption on Wikipedia and other Web platforms.

Media Bias in German Online Newspapers

Online newspapers have been established as a crucial information source, at least partially replacing traditional media like television or print media. As all other media, online newspapers are potentially affected by media bias.This describes non-neutral reporting of journalists and other news producers, e.g. with respect to specific opinions or political parties. Analysis of media bias has a long tradition in political science. However, traditional techniques rely heavily on manual annotation and are thus often limited to the analysis of small sets of articles. In this paper, we investigate a dataset that covers all political and economical news from four leading German online newspapers over a timespan of four years. In order to analyze this large document set and compare the political orientation of different newspapers, we propose a variety of automatically computable measures that can indicate media bias. As a result, statistically significant differences in the reporting about specific parties can be detected between the analyzed online newspapers.

Difference-based estimates for generalization-aware subgroup discovery

For the task of subgroup discovery, generalization-aware interesting measures that are based not only on the statistics of the patterns itself, but also on the statistics of their generalizations have recently been shown to be essential. A key technique to increase runtime performance of subgroup discovery algorithms is the application of optimistic estimates to limit the search space size. These are upper bounds for the interestingness that any specialization of the currently evaluated pattern may have. Until now these estimates are based on the anti-monotonicity of instances, which are covered by the current pattern. This neglects important properties of generalizations. Therefore, we present in this paper a new scheme of deriving optimistic estimates for generalization aware subgroup discovery, which is based on the instances by which patterns differ in comparison to their generalizations. We show, how this technique can be applied for the most popular interestingness measures for binary as well as for numeric target concepts. The novel bounds are incorporated in an efficient algorithm, which outperforms previous methods by up to an order of magnitude.

Taking OWL to athens: semantic web technology takes ancient greek history to students

The HermesWiki project is a semantic wiki application on Ancient Greek History. As an e-learning platform, it aims at providing students effective access to concise and reliable domain knowledge, that is especially important for exam preparation. In this paper, we show how semantic technologies introduce new methods of learning by supporting teachers in the creation of contents and students in the personalized identification of required knowledge. Therefore, we give an overview of the project and characterize the semi-formalized content. Additionally, we present several use cases and describe the semantic web techniques that are used to support the application. Furthermore, we report on the user experiences regarding the usefulness and applicability of semantic technologies in this context.