Research Article

A Performance Comparison of Unsupervised Techniques for Event Detection from Oscar Tweets

Figure 5

A graphical representation of the Oscars dataset using K-SVD for 5, 10, 15, and 20 clusters. When k = 5 and k = 10, the generated clusters are distanced from one another and distinct; however, when k = 15 and k = 20, the formed clusters are overlapped. (a) Clusters = 5. (b) Clusters = 10. (c) Clusters = 15. (d) Clusters = 20.
(a)
(b)
(c)
(d)