Mapping Evaluation for Semantic Browsing

Abstract

Abstract-The paper contributes to the problem solving in semantic browsing and analysis of scientific articles. With reference to presented visual interface, four -the most popular methods of mapping including own approach -MDS with spherical topology, have been compared. For a comparison quantitative measures were applied which allowed to select the most appropriate mapping way with an accurate reflection of the dynamics of data. For the quantitative analysis the authors used machine learning and pattern recognition algorithms and described: clusterization degree, fractal dimension and lacunarity. Local density differences, clusterization, homogeneity, and gappiness were measured to show the most acceptable layout for an analysis, perception and exploration processes. Visual interface for analysis how computer science evolved through the two last decades is presented on website. Results of both quantitative and qualitative analysis have revealed good convergence. I. INTRODUCTIONonlinear growth of scientific writing imposes a new forms of academic databases management. The latter includes the both retrieval and analytical exploration. Analysts, science of science professionals, science policy makers need various computing, statistical and visualization tools to monitor how science or scientific domains evolved. N Authors designed and described in a series of papers [1], [2] the visual interface for analysis of dynamics of computer science through the two last decades 1 . Screenshot of spherical application is shown on Figure 1. Users can interact, manipulate and browse the data and see how graphical pattern change in time. The nodes represent scientific articles from digital library and the colour -appropriate thematic category. Similarity metrics was based on semantic relations between documents [2]. In order to generate 3D layout, multidimensional scaling (MDS) technique was applied and enriched by Morse potential [1].Overlapping spots show where the categories mutually integrate that means an articles at that location are semantically similar. Visualization of classified documents reveals both organization of digital library content as well as allows users to track how it changes over time. This paper presents fur-1 http://www-users.mat.umk. pl/~garfi/vis2009v3/ ther study on visualization maps. Authors decided to test this prototype regarding to mapping algorithms. Four mapping methods were compared in terms of dynamics and analytical possibility of output visualization. The next chapter shows the outline of VxOrd, MDS, VOS, SOM as the most popular methods. II.MAPPING PREVIEWHow we, as analysts, perceive and understand the connections between data, depends on graphical layout. Thus, the final structure of visualized knowledge can be drawn either by spatial arrangement (2D or 3D) of analysis units or by the relationship between nodes in graph or combination of these two.One of the basic ordination algorithm -VxOrd extends a traditional force-directed approach [3]. VxOrd determines the both number and size of ...

Game with a Purpose for Mappings Verification

“…The results often very complicated algorithms are quite easy to grasp by an average human. Linking those two areas could prove to be useful for performing laborious tasks without a need of hiring additional workers especially that many research results needs some evaluation and sometimes it is the sole purpose of the research [6].…”