Computer Science and Languages

A hierarchical structure can provide efficient access to information contained in a collection of documents. However, such a structure is not always available, e.g. for a set of documents a user has collected over time in a single folder or the results of a web search. We therefore investigate in this paper how we can obtain a hierarchical structure automatically, taking into account some background knowledge about the way a specific user would structure the collection. More specifically, we adapt a hierarchical agglomerative clustering algorithm to take into account user specific constraints on the clustering process. Such an algorithm could be applied, e.g., for user specific clustering of Web search results, where the user's constraints on the clustering process are given by a hierarchical folder or bookmark structure. Besides the discussion of the algorithm itself, we motivate application scenarios and present an evaluation of the proposed algorithm on benchmark data.

Although clustering under constraints is a current research topic, a hierarchical setting, in which a hierarchy of clusters is the goal, is usually not considered. This paper tries to fill this gap by analyzing a scenario, where constraints are derived from a hierarchy that is partially known in advance. This scenario can be found, e.g., when structuring a collection of documents according to a user specific hierarchy. Major issues of current approaches to constraint based clustering are discussed, especially towards the hierarchical setting. We introduce the concept of hierarchical constraints and continue by presenting and evaluating two approaches using them. The approaches cover the two major fields of constraint based clustering, i.e. instance and metric based constraint integration. Our objects of interest are text documents. Therefore, the presented algorithms are especially fitted to work for these where necessary. Despite showing the properties and ideas of the algorithms in general, we evaluated the case of constraints that are unevenly scattered over the instance space, which is very common for real-world problems but not satisfyingly covered in other work so far.

In order to organize huge document collections, labeled hierarchical structures are used frequently. Users are most efficient in navigating such hierarchies, if they reflect their personal interests. Thus, we propose in this article an approach that is able to derive a personalized hierarchical structure from a document collection. The approach is based on a semi-supervised hierarchical clustering approach, which is combined with a biased cluster extraction process. Furthermore, we label the clusters for efficient navigation. Besides the algorithms itself, we describe an evaluation of our approach using benchmark datasets.

Hierarchical classification refers to an extension of the standard classification problem, in which labels must be chosen from a class hierarchy. In this paper, we look at hierarchical classification from an information retrieval point of view. More specifically, we consider a scenario in which a user searches a document in a topic hierarchy. This scenario gives rise to the problem of predicting an optimal entry point, that is, a topic node in which the user starts searching. The usefulness of a corresponding prediction strongly depends on the search behavior of the user, which becomes relevant if the document is not immediately found in the predicted node. Typically, users tend to browse the hierarchy in a top-down manner, i.e., they look at a few more specific subcategories but usually refuse exploring completely different branches of the search tree. From a classification point of view, this means that a prediction should be evaluated, not solely on the basis of its correctness, but rather by judging its usefulness against the background of the user behavior. The idea of this paper is to formalize hierarchical classification within a decision-theoretic framework which allows for modeling this usefulness in terms of a user-specific utility function. The prediction problem thus becomes a problem of expected utility maximization. Apart from its theoretical appeal, we provide first empirical results showing that the approach performs well in practice.

In order to navigate huge document collections efficiently, tagged hierarchical structures can be used. For users, it is important to correctly interpret tag combinations. In this paper, we propose the usage of tag groups for addressing this issue and an algorithm that is able to extract these automatically for text documents. The approach is based on the diversity of content in a document collection. For evaluation, we use methods from ontology evaluation and showed the validity of our approach on a benchmark dataset.

In folk song research, appropriate similarity measures can be of great help, e.g. for classification of new tunes. Several measures have been developed so far. However, a particular musicological way of classifying songs is usually not directly reflected by just a single one of these measures. We show how a weighted linear combination of different basic similarity measures can be automatically adapted to a specific retrieval task by learning this metric based on a special type of constraints. Further, we describe how these constraints are derived from information provided by experts. In experiments on a folk song database, we show that the proposed approach outperforms the underlying basic similarity measures and study the effect of different levels of adaptation on the performance of the retrieval system.

In order to support the navigation in huge document collections efficiently, tagged hierarchical structures can be used. Often, multiple tags are used to describe resources. For users, it is important to correctly interpret such tag combinations. In this paper, we propose the usage of tag groups for addressing this issue and an algorithm that is able to extract these automatically for text documents. The approach is based on the diversity of content in a document collection and a user's structuring preference. For evaluation, we use methods from ontology evaluation and show the validity of our approach on a benchmark dataset.

Constrained clustering has many useful applications. In this paper, we consider applications, in which a hierarchical target structure is preferable. Therefore, we constrain a hierarchical agglomerative clustering through the use of MLB constraints, which provide information about hierarchical relations between objects. We propose an algorithm that learns a suitable metric according to the constraint set by modifying the metric whenever the clustering process violates a constraint. We furthermore combine this approach with an instance-based constrained clustering to further improve the cluster quality. Both approaches have proven to be very successful in a semi-supervised setting, in which constraints do not cover all existing clusters.

Keeping one's personal music collections well organized can be a very tedious task. Fortunately, today, many popular music players (such as AmaroK or iTunes) have an integrated library function that can automatically rename and tag music files and sort them into subdirectories. However, their common approach to stick with some hierarchy of genre, artist name, and album title barely represents the way a user would structure his collection manually. When it comes to organizing a music collection according to a user-specific hierarchy, three things are required: First, the music files have to be described by appropriate features beyond simple meta-tags. This includes content-based analysis but also incorporation of external information sources such as the web. Second, knowledge about the user's structuring preferences must be available. And third, and most importantly, methods for learning personalized hierarchies that can integrate this knowledge are needed. We propose for this task a hierarchical constraint based clustering approach that can weight the importance of different features according to the user perceived similarity. A hierarchy that is built based on this similarity measure reflects a user's view on the collection.

In this paper we describe the development of a user interface for a mobile search application. The interface is connected to an information retrieval system in order to allow mobile searching and personal information organization by using clustering algorithms. By making use of novel interaction techniques like multi-touch, users are supported in refining their information need. In addition users are provided with a tag cloud of past queries in order to offer a personalized initialization of the search process. We have conducted several user studies during the entire development process. In order to follow a formative evaluation process also a final usability test was conducted, which reveals a high usability of the concept.

Information management is a tedious task that we wish to support through automatic methods. In the past, we developed different data mining techniques capable of providing automatic structuring to be used in information management. In this paper, we discuss how those algorithms can be employed in a system assisting the user through appropriate visualization and interaction. We identify requirements on the system, propose several design elements and describe how they enable effective interaction between the user and the data mining methods. Those form the base for future user studies.

In this paper we describe the development of a user interface for a mobile search application. The interface is connected to an information retrieval system in order to allow mobile searching and personal information organization by using clustering algorithms. By making use of novel interaction techniques like multi-touch, users are supported in refining their information need. In addition users are provided with a tag cloud of past queries in order to offer a personalized initialization of the search process. We have conducted several user studies during the entire development process. In order to follow a formative evaluation process also a final usability test was conducted, which reveals a high usability of the concept.

In this paper we describe the results of a study which compares explicit synchronous collaborative web search with searching the web individually when users are utilizing the web for the purpose of social planning tasks, in specific travel planning. We explore both co-located and distributed collaborative search within groups of two and three participants. The distributed collaborative search is supported by SearchTogether. The results show that distributed collaborative search in general outperforms both co-located and individual search. Surprisingly, the increase of the group size negatively influences the search results, though it positively affects the user confidence in the search results. We conclude that the negative effect arises due to an increased communication effort and suggest future work in HCI to better support collaboration in large groups.

Constrained clustering received a lot of attention in the last years. However, the widely used pairwise constraints are not generally applicable for hierarchical clustering, where the goal is to derive a cluster hierarchy instead of a flat partition. Therefore, we propose for the hierarchical setting-based on the ideas of pairwise constraints-the use of must-link-before (MLB) constraints. In this paper, we discuss their properties and present an algorithm that is able to create a hierarchy by considering these constraints directly. Furthermore, we propose an efficient data structure for its implementation and evaluate its effectiveness with different datasets in a text clustering scenario.

Unpersonalized web search has come to its limits as the size of the World Wide Webisdaily increasing and search results for manytopics contain hundreds or even thousands of documents. In this paper,wepresent an approach for personalizing web search using the bookmarks of au ser.T his information is used to classify the search results into the bookmark hierarchy of au ser.A sw ea re dealing with ac lass hierarchyinstead of a fl at class structure, the classifi cation task changes compared to standard text classifi cation. We are presenting twoa lgorithms solving the hierarchical classi fi cation task and evaluate them using asuited performance measure close to standard precision and recall.

This paper presents results of the research project “partiMAN – the Participation Manager for Collaborative Decision Making”. These include practical requirements for the design of planning and decision making processes. A wide research of the status quo of regional planning leads to future important factors that support e-participation. Besides an improved data acquisition and data analysis of user-related content, new ways shall be sought to connect already existing as well as newly created factual and geographic data. Related results shall properly be analyzed and published.

The planning of built and natural environments incorporates participatory processes that are more and more supported by information and communications technology (ICT). On the one hand, this implies potentially more diverse knowledge of the masses in form of participatory contributions, e. g., opinions, ideas or formal statements. On the other hand, this leads to an increasing data complexity in terms of size and interconnectedness which in turn complicates making sense, retrieving implicit new information, and decision making. We describe the need for more sophisticated methods and present how the field of visual analytics can overcome these challenges by demonstrating the applicability of the visual analytics process model for one exemplary but fundamental use case that focuses on an analysis task in e-participation for planning purposes.