In this project the marginal utility of Network Graph Analysis is explored in a text analysis setting. How does the value-add stack up? Nuclear.... or Neutral.
Conventional and more rudimentary techniques such as Term Freqeuncy Analysis, Term Weighting, Topic Modelling and Cluster Analysis delivered modest insights. However, when Netwrok Graph Analysis was combined with Latent Semantic techniques the door to much richer insights began to swing open.
The following images are two page extracts from the detailed report which is avaialble for download at the end of this post and which hints at some of the more potent insights that can flow from the more sophisticated approaches.
Download the full report:
Comments