solve questions

As a sccond explotation, using the complete and/or consmonalities among the three joumals and on data set (incloding all three joumals and all four potentlal changes in the einphasis on thone chareers: periods), Delen and Crossland conducted a cliaster that is, to answer questions such as "Are there clusing analysis. Chustering is arguably the most com - ters that represent different reveatch themes specific monly tised text mining lechanique., Clustering was to a single fournap and "ts there a time-varying tised in this study to identify the natural groupings characterization of those clusteri The sevearchen of the articles (by putting them into separate clus- discovered and divcussed several interesting patters) and then to list the most descriptive terms tems using tabular and graphical representation of that chanicterixed those clusters. They used SVD to their findings (for further information, see Delen recluce the dinaensionality of the term-by-document and Crosiland, 2008). matrix and then an expectation-maximization algorithm to create the clusters. They conducted sevcral experiments to identify the optinsal number Quzsmoss roue Case 7.5 of clusters, which tumest out to be ninc, After the 1. How can text mining be uscd to case the insurconstruction of the nine clusters, they analyzed the nountable task of literature review? content of those clasters from rwo perspectives: 2. What are the common outcomes of a text mining (1) representation of the journal type (see Figure project on a specific collection of journal articles? 7.8a ) and (2) representation of time (Figure 7.8b). Can you think of other posential outcomes nox The idea was to explore the potential differcnces mentioned in this case