Trend analysis based on semantic graphs - A case study

Abstract. Information published on the web is continuously growing and getting more and more complex. It is possible to accomplish quantitative trend analysis on collected data in various aspects. This paper describes a methodology in order to explore tendencies targeting the cloud and the quality characteristics (especially the testing domain) inside. The implemented prototype does frequency analysis on well prepared data sets of academic papers published on the web. The results are represented with a sequence of semantic graphs. This kind of time series data is an extension of the traditional tag cloud definitions. The paper describes the developed web crawling technology as well. The presented method can be applied as a reusable design pattern for any kind of trend analysis based on web data.