The Influence Of Font Scale On Semantic Expression Effects Of Word Cloud

Since we started to contact with the field of visualization,we know that an important purpose of visualization research is to present data through various visual methods so that people can better understand and analyze the data.Word clouds are often used as an efficient and necessary method in the study of visualization of text data.In many real-world applications,a commonly used method is using a word cloud to reduce the difficulty for people to understand,grasp the main idea of document quickly and reduce the time spent,so as to better express the semantic understanding of a document.Therefore,many researchers have devoted a lot of time and energy to studying word clouds,including the algorithm of word cloud layout,and the attributes of word cloud such as shape,font,color,word spacing and so on.In the process of generating word clouds from documents,it is a common usage that the font sizes of words are proportional to their frequencies in the document.Consequently,one important problem with word cloud is how to set word font sizes to facilitate understanding of semantics at the best situation.In this paper,we will explore this question with a controlled experiment.First of all,in this paper,we proposed a novel perspective of attribute to describe the word clouds,called the relative font size.Through experimental design,we explored the influence of the relative font size(in this paper,called scale)on people’s understanding of word cloud semantics from a overall and global perspective.The challenge is how to establish a rigorous experiment environment that enables accurate quantification of rich semantic information involved in a document.For this purpose,we utilize an LDA ensemble-based method to support interactive selection of interesting topics with the purpose of avoiding the influence of the model parameters to the extracted topics.We conducted our study through two pilot experiments and two formal experiments.Two pilot experiments determine the perspective and direction of experimental research in the formal experiment.Through formal experiment one,we found that different scale of word cloud would indeed affect people’s understanding of semantics according to the accuracy rate,completion time and completion confidence.In the formal experiment two,we refined and classified the experimental results and conducted a more detailed study to explore the different influence patterns of scale on semantic understanding of word clouds due to the differences between word clouds.Finally,according to the experimental results,we put forward reasonable suggestions for readers to achieve better semantic understanding of word cloud.Our findings aimed at optimizing the scale of word cloud and improving its semantic expressing ability.