Table 2 Examples of randomly chosen clusters with randomly chosen k-values (k of k-means clustering)
From: Assessing the role of participants in evolution of topic lifecycles on social networks
k-value | Cluster content |
|---|---|
Example clusters from June 2009 data | |
500 | #trends, #trendy, #trend, #fashion |
400 | #icon, #Revolution, #freedom, #election, #revolution, #world |
300 | #Cricket, #PakCricket, #T20, #pakcricket, #cricket, #t20 |
200 | #recipes, #food, #gfree, #cooking, #nom, #primal, #foodie, #vegan, #recipe |
100 | #wedding, #vintage, #etsy, #flowers, #makeup, #jewelry, #gardening, #zazzle, #artfire, #twittographers, #beauty, #giveaway, #handmade, #Etsy, #shoes, #etsytwitter |
Example clusters from July 2009 data | |
2000 | #Swineflu, #pandemic |
1600 | #Weather, #weather, #Denmark, #lightning, #Copenhagen |
1200 | #mashup, #remix, #soul, #HipHop, #mv, #chill |
800 | #Ford, #automotive, #Cars, #car, #Remart, #ford, #hybrid, #used, #Car, #Autoparts |
400 | #demilovatolive, #DLovatoOnUstream, #DisneyRecords, #JBsouthamerica, #DemiLovatoLive, #DemiLiveWebcast |