%0 Conference Paper %1 rodrigues2017vaforwordrelevances %A Rodrigues, Nils %A Burch, Michael %A Di Silvestro, Lorenzo %A Weiskopf, Daniel %B Proceedings of the International Conference on Information Visualisation (IV) %D 2017 %I IEEE %K 2017 B01 from:mueller sfbtrr161 vis(us) visus visus:burchml visus:rodrigns visus:weiskopf %P 1-7 %R 10.1109/iV.2017.62 %T A Visual Analytics Approach for Word Relevances in Multiple Texts %U https://ieeexplore.ieee.org/document/8107940 %X We investigate the problem of analyzing word frequencies in multiple text sources with the aim to give an overview of word-based similarities in several texts as a starting point for further analysis. To reach this goal, we designed a visual analytics approach composed of typical stages and processes, combining algorithmic analysis, visualization techniques, the human users with their perceptual abilities, as well as interaction methods for both the data analysis and the visualization component. By our algorithmic analysis, we first generate a multivariate dataset where words build the cases and the individual text sources the attributes. Real-valued relevances express the significances of each word in each of the text sources. From the visualization perspective, we describe how this multivariate dataset can be visualized to generate, confirm, rebuild, refine, or reject hypotheses with the goal to derive meaning, knowledge, and insights from several text sources. We discuss benefits and drawbacks of the visualization approaches when analyzing word relevances in multiple texts.

@inproceedings{rodrigues2017vaforwordrelevances, abstract = {We investigate the problem of analyzing word frequencies in multiple text sources with the aim to give an overview of word-based similarities in several texts as a starting point for further analysis. To reach this goal, we designed a visual analytics approach composed of typical stages and processes, combining algorithmic analysis, visualization techniques, the human users with their perceptual abilities, as well as interaction methods for both the data analysis and the visualization component. By our algorithmic analysis, we first generate a multivariate dataset where words build the cases and the individual text sources the attributes. Real-valued relevances express the significances of each word in each of the text sources. From the visualization perspective, we describe how this multivariate dataset can be visualized to generate, confirm, rebuild, refine, or reject hypotheses with the goal to derive meaning, knowledge, and insights from several text sources. We discuss benefits and drawbacks of the visualization approaches when analyzing word relevances in multiple texts.}, added-at = {2020-10-09T12:34:20.000+0200}, author = {Rodrigues, Nils and Burch, Michael and Di Silvestro, Lorenzo and Weiskopf, Daniel}, biburl = {https://puma.ub.uni-stuttgart.de/bibtex/240fb11b76702dc45453f568e3d29a339/mueller}, booktitle = {Proceedings of the International Conference on Information Visualisation (IV)}, description = {A Visual Analytics Approach for Word Relevances in Multiple Texts}, doi = {10.1109/iV.2017.62}, interhash = {a61be20ea953ab79ea5a5998feca0911}, intrahash = {40fb11b76702dc45453f568e3d29a339}, keywords = {2017 B01 from:mueller sfbtrr161 vis(us) visus visus:burchml visus:rodrigns visus:weiskopf}, pages = {1-7}, publisher = {IEEE}, timestamp = {2020-10-09T10:34:20.000+0200}, title = {A Visual Analytics Approach for Word Relevances in Multiple Texts}, url = {https://ieeexplore.ieee.org/document/8107940}, year = 2017 }