Results 1 - 10 of 62,639

Table 1: Clustering ensemble and consensus solution

in Abstract A Mixture Model for Clustering Ensembles
by unknown authors
"... In PAGE 7: ... Correspondence problem is emphasized by different label systems used by the partitions. Table1 shows the expected values of latent variables after 6 iterations of the EM algorithm and the resulting consensus clustering. In fact, stable combination appears even after the third iteration, and it corresponds to the true underlying structure of the data.... In PAGE 10: ... Figure 3 shows the error as a function of k for different consensus functions for the galaxy data. It is also interesting to note that, as expected, the average error of consensus clustering was lower than average error of the k-means clusterings in the ensemble ( Table1 ) when k is chosen to be equal to the true number of clusters. Moreover, the clustering error obtained by EM and MCLA algorithms with k=4 for Biochemistry data was the same as found by the advanced supervised classifiers applied to this dataset [28].... ..."

Table 3. Mean running times (in seconds) for ensemble clustering procedures.

in Efficient Ensemble Methods for Document Clustering
by Derek Greene, Pádraig Cunningham
"... In PAGE 10: ...3 Comparison of Algorithm Efficiency Another important aspect of our evaluation was to assess the computational gains resulting from prototype reduction. Table3 provides a list of the mean running times for the ensemble clustering experiments, which were performed on a Pentium IV 3.... ..."

Table 2: Classi cation error on MNIST training, validation, and test sets, with the best hyper- parameters according to validation error, with and without pre-training, using purely supervised or purely unsupervised pre-training. In experiment 3, the size of the top hidden layer was set to 20. On MNIST, differences of more than .1% are statistically signi cant. The results in Table 2 suggest that the auto-encoding criterion can yield performance comparable to the DBN when the layers are nally tuned in a supervised fashion. They also clearly show that the greedy unsupervised layer-wise pre-training gives much better results than the standard way to train a deep network (with no greedy

in Greedy layer-wise training of deep networks
by Yoshua Bengio, Pascal Lamblin, Dan Popovici, Hugo Larochelle 2006
"... In PAGE 7: ... A possible expla- nation is that the greedy supervised procedure is too greedy: in the learned hidden units representation it may discard some of the information about the target, information that cannot be captured easily by a one-hidden-layer neural network but could be captured by composing more hidden layers. Experiment 3 However, there is something troubling in the Experiment 2 results ( Table2 ): all the networks, even those without greedy layer-wise pre-training, perform almost perfectly on the training set, which would appear to contradict the hypothesis that the main effect of the layer-wise greedy strategy is to help the optimization (with poor optimization one would expect poor training error). A possible explanation coherent with our initial hypothesis and with the above results is captured by the following hypothesis.... In PAGE 7: ... To test that hypothesis, we performed a second series of experiments in which we constrain the top hidden layer to be small (20 hidden units). The Experiment 3 results ( Table2 ) clearly con rm our hypothesis. With no pre-training, training error degrades signi cantly when there are only 20 hidden units in the top hidden layer.... ..."
Cited by 11

Table 2: Function names of signals in clusters

in Linear Modeling of Genetic Networks from Experimental Data
by E.P. van Someren, L.F.A. Wessels, M.J.T. Reinders
"... In PAGE 9: ... 11, has disappeared and is more sparse than the matrix obtained with mean vari- ance normalization. Table2 lists the function names associated with the genes in the clusters. There are two dominant clusters, namely, Cluster 2 which con- sists mainly of genes involved in MATING and Cluster 7 which consists entirely of CHROMATIN STRUCTURE related genes.... ..."

Table 2 Cluster Reconstruction

in Energy Efficient Multipath Routing in Large Scale Sensor Networks with Multiple Sink Nodes
by Yuequan Chen, Edward Chan, Song Han

Table 4. The geometrical mean percentage of clusterers selected by selective voting and selective weighted-voting under different ensemble sizes. ensemble size percentage of selecting

in Abstract Clusterer Ensemble
by Zhi-hua Zhou, Wei Tang

Table 5. Embedded shifter modes and control signals

in TABLE OF CONTENTS
by Michael J. Beauchamp, Michael J. Beauchamp, Professor W. H. Carl Ebeling 2006
"... In PAGE 35: ... The 66 outputs include 64 data bits and 2 sticky bits (two independent sticky bit outputs are needed when the shifter is used as two independent 32-bit shifters). The modes of operation and control of the embedded shifter is given in Table5 . A 64-bit left and right shifter can be created out of a 127-bit right only shifter.... In PAGE 35: ... The bottom five levels of muxes need to be duplicated and the control for the top level was modified to account for both 32-bit and 64-bit shift modes. The control for the embedded shifter is shown in Table5 . The control was determined such that in both the 32-bit and 64-bit shift modes the output is in correct sequence and another level muxes are not needed.... ..."

Table 2: SNR of watermark embedded audio signal

in Digital Piracy on P2P Network Summary
by unknown authors
"... In PAGE 5: ... Since the unexpected number of zeros is added at the front of audio signal during compression, a preliminary work to obtain the information about the start position of watermarking was done. The experimental results are given in Table2 which shows that watermark using the 2D barcode is not affected by down-sampling and lossy-compression. Good error- correcting with 2D barcode itself can enhance watermark detection.... ..."

Table 2 Summary of the design of the 7 ensembles types Number

in Moderate diversity for better cluster ensembles
by Stefan T. Hadjitodorov A, Ludmila I. Kuncheva B, Ludmila P. Todorova A 2005
"... In PAGE 6: ...ig. 4. Fitted polynomial of degree 3 for the ensemble accuracy versus averaged individual accuracy (dot marker). structed as summarized in Table2 . Two most common types of clusterers were used: the k-means and the mean link method (average link, average linkage).... ..."

Table 5: Reconstruction error on entire ERIM

in A framework for simultaneous co-clustering and learning from complex data
by Meghana Deodhar, Joydeep Ghosh 2007
"... In PAGE 8: ...2.3 Data Reconstruction Table5 displays the mean squared error of the approx- imated data matrix obtained by the different algorithms on the entire standardized dataset, averaged over 10 runs. Row Clustering is Model CC with the number of column clusters set to 1 and Column Clustering is Model CC with the number of row clusters set to 1.... In PAGE 8: ... Note that Model CC obtains the best reconstruction of the original matrix as compared to the other approaches in terms of MSE (mean squared error). Table5 also shows the average R2 of the linear models constructed within each co-cluster. The R2 values are actually quite low, indicating that a strong lin- ear relationship does not really exist in the data, which is to the disadvantage of the simultaneous co-clustering and regression algorithm.... ..."
Cited by 1
Results 1 - 10 of 62,639