Social Media Data Mining and Analytics

Index

• 80/20 rule (p/q rule), 28–32

• A
  • activities
    • burst activity, 99
    • fraction of most active users, 28–29
    • graphs from, 48–51
    • patterns, 293–294
  • activity buckets, 7, 16
    • logarithmic binning, 36
    • revision buckets, 12
    • time, number of Tweets, 82–83
  • agglomerative clustering, 130–132
  • aggregate metrics, streaming and, 57
  • algorithms
    • approximative, 216
    • breadth-first search, 51
    • depth-first search, 51
    • graphs and, 44
    • hierarchical agglomerative algorithm, 140
    • sLDA, 156–162
  • Amazon EC2, CDH distribution, 237–243
  • APIs (application programming interfaces), Twitter, 32
  • approximations, 214–236
  • ARIMA (Autoregressive Integrated Moving Average), 117–121
  • assortativity, 70–75
  • asymmetrical social network, 56
  • AUC (area under the curve), 282
  • autocorrelations, 89–91
  • averages, 84–85, 296–303

• B
  • Bayesian nonparametrics, 262
  • BeautifulSoup, 126–127
  • BF (Bloom filter) algorithm, 216, 223–231
  • buckets. See activity buckets
  • burst activity, 99–101, 105–109

• C
  • cascading, 177
  • check-ins, 249
  • Cloudera CDH distribution (Hadoop), 237–243
  • clustering
    • complete linkage, 131
    • content topics, 130
    • dendrogram, 135–137, 140–141
    • hierarchical agglomerative clustering, 130–131
    • k-means, 199–203
    • links, 64–65
    • local clustering coefficient, 64–66
    • merging heights, 135, 137
    • on-demand, CDH distribution and, 242–243
    • single linkage, 131
  • CMS (Count-Min Sketch) algorithm, 216, 231–236
  • cold-start problem, recommendations and, 245
  • collaborative filtering, 48
  • commutative monoids, 216–217
  • complementary cumulative ...