|Math Coffee: "A Tale of Four Flowers" Daniel Martin '11
When: November 04,
Speaker: Daniel Martin '11Ticket Required: No
This summer, I investigated the fields of data mining and cluster analysis at the NC State Applied Math REU. Clustering is the act of assigning a set of objects into subsets, or clusters, so that objects in the same cluster are similar. We began our research into this area by learning about several known clustering techniques such as k-means and hierarchical clustering. Using two sets of documents, the REU paragraphs and the Blackstone documents, we created a search engine using latent semantic indexing and clustered the documents using non-negative matrix factorization. We also used two spectral techniques, the Fiedler Method and the MinMaxCut Method, to cluster these documents as well as Fisher's Iris data set. I will discuss these results as well as the overall REU experience.
Gather in Mathematics Hall at 6:30 p.m. for traditional light snacks.
Contact: Prof. Tim Chartier