{"responseHeader":{"status":0,"QTime":7,"params":{"q":"{!q.op=AND}id:\"102836\"","hl":"true","hl.simple.post":"","hl.fragsize":"5000","fq":"!embargo_tdt:[NOW TO *]","hl.fl":"ocr_t","hl.method":"unified","wt":"json","hl.simple.pre":""}},"response":{"numFound":1,"start":0,"docs":[{"file_name_t":"Soller-Automated_Detection.pdf","thumb_s":"/9b/7c/9b7c1044eaaa76c5b9eed10d174ca1a59e88fff2.jpg","oldid_t":"compsci 10949","setname_s":"ir_computersa","restricted_i":0,"format_t":"application/pdf","modified_tdt":"2016-05-26T00:00:00Z","file_s":"/0f/26/0f266384fcd924663b7715e6f540294d576ede0d.pdf","title_t":"Page 79","ocr_t":"64 10.3.3 Tradeoffs The possible advantages of RBF networks over the multilayer perceptron include • computationally faster convergence of learning algorithm; • improved approximation in some problems; and • more conceptually meaningful solutions through interpretable clusters. The possible disadvantages of RBF networks include • poorer generalization in some problems; and • poorer approximation in some problems. 10.4 Winner Take All Networks In the radial basis function networks, an input can produce nonzero threshold function outputs from more than one cluster. In some situations, it is desirable to only characterize an input by its cluster membership and not by its Euclidean distances to various cluster centers. This dissertation calls this architecture the winner take all network. Each cluster is determined in an unsupervised k-means clustering fashion, as discussed above. After the centers are determined, each cluster is labelled with a classification. This architecture is a limiting case of the radial basis function network; it is equivalent to a radial basis function network which uses small widths and large positive or negative weights on the second layer. Although supervised learning can classify the centers, this project chose a more direct Bayesian classification method where z · p(cl(x)lx E 8i) = 2 t · J • cl ( ·) represents the correct classification of delirium; • clusters are classified by their majority class; (10.30)","id":102836,"created_tdt":"2016-05-26T00:00:00Z","parent_i":102961,"_version_":1642982670131527681}]},"highlighting":{"102836":{"ocr_t":[]}}}