Search

Search Funnelback University

Search powered by Funnelback
51 - 70 of 314 search results for Economics test |u:mlg.eng.cam.ac.uk where 23 match all words and 291 match some words.
  1. Results that match 1 of 2 words

  2. Sparse Gaussian Processes using Pseudo-inputs Edward Snelson Zoubin…

    https://mlg.eng.cam.ac.uk/zoubin/papers/nips05spgp.pdf
    27 Jan 2023: Once the inversion is done, prediction isO(N ) for thepredictive mean andO(N 2) for the predictive variance per new test case. ... We have demonstrated a significant decrease in test error over the other methods for a givensmall pseudo/active set size.
  3. AA06.dvi

    https://mlg.eng.cam.ac.uk/pub/pdf/GirRasQuiMur03.pdf
    13 Feb 2023: the density of the actual true test output under the Gaussianpredictive distribution and use its negative log as a measure of loss. ... The training and test data consist of pH values (outputsy of the process) anda control input signal (u).
  4. Large Scale Nonparametric Bayesian Inference:Data Parallelisation in…

    https://mlg.eng.cam.ac.uk/pub/pdf/DosKnoMohGha09.pdf
    13 Feb 2023: Initially, a largenumber of features are added, which provides improvements in the test likelihood. ... Table 2 summarises the data and shows thatall approaches had similar test-likelihood performance.
  5. paper.dvi

    https://mlg.eng.cam.ac.uk/pub/pdf/KimGha08.pdf
    13 Feb 2023: Figure 1 shows the training points, test points, decision boundary from GPCand decision boundary from robust GPC. ... We created 10 pairs of training and test sets by randomlydividing the whole data set into two.
  6. Bayesian Classifier Combination Zoubin Ghahramani and Hyun-Chul Kim∗…

    https://mlg.eng.cam.ac.uk/zoubin/papers/GhaKim03.pdf
    27 Jan 2023: Satellitehas a training set of 4435, a test set of 2000 with 6 classes and 36 variables. ... UCI digit data set has a trainingset of 3823, a test set of 1797, 10 classes and 64 variables.
  7. - 4F13: Machine Learning

    https://mlg.eng.cam.ac.uk/teaching/4f13/0708/lect06.pdf
    19 Nov 2023: Constraint-Based Learning: Use statistical tests of marginal and conditionalindependence. Find the set of DAGs whose d-separation relations match theresults of conditional independence tests.
  8. Manifold Gaussian Processes for Regression Roberto Calandra∗, Jan…

    https://mlg.eng.cam.ac.uk/pub/pdf/CalPetRasDei16.pdf
    13 Feb 2023: Additionally, for the test set, wemake use of the Negative Log Predictive Probability (NLPP). ... For training we extract 400consecutive data points, while we test on the following 500data points.
  9. LNCS 5342 - Outlier Robust Gaussian Process Classification

    https://mlg.eng.cam.ac.uk/pub/pdf/KimGha08a.pdf
    13 Feb 2023: Figure 1 shows the training points, test points, decision boundary from GPCand decision boundary from robust GPC. ... We created 10 pairs of training and test sets by randomlydividing the whole data set into two.
  10. Randomized Nonlinear Component Analysis

    https://mlg.eng.cam.ac.uk/pub/pdf/LopSraSmo14a.pdf
    13 Feb 2023: Results are statistically significantrespect to a paired Wilcoxon test on a 95% confidence inter-val. ... Figure 3. Autoencoder reconstructions of unseen test images forthe MNIST (top) and CIFAR-10 (bottom) datasets.
  11. Local and global sparse Gaussian process approximations Edward…

    https://mlg.eng.cam.ac.uk/zoubin/papers/aistats07localGP.pdf
    27 Jan 2023: The nearestblock’s GP is used to predict at a given test point. ... At test time, a test point is simply assigned tothe nearest cluster center.
  12. 13 Feb 2023: The test data is created by randomlyselecting 10% of the data points and setting them as missing. ... 0.5. 1. 1.5. 2. 2.5. 3. PARAFAC. Probabilistic NTF. Test Train.
  13. BIOINFORMATICS ORIGINAL PAPER Vol. 21 no. 16 2005, pages ...

    https://mlg.eng.cam.ac.uk/pub/pdf/ChuGhaFal05a.pdf
    13 Feb 2023: Non-parametric tests,e.g. the Wilcoxon rank sum test, are superior to the t -test in this case. ... The integers in the parantheses is the total test error numberover the 10 folds.
  14. /users/joe/src/tops/dvips

    https://mlg.eng.cam.ac.uk/pub/pdf/UedNakGha00a.pdf
    13 Feb 2023: improve the likelihood of both the training dataand of held-out test data. ... The split criterion defined by equation 3.13 can be viewed as a likelihoodratio test.
  15. Infinite Hidden Markov Models and extensions

    https://mlg.eng.cam.ac.uk/zoubin/talks/BayesHMMs10.pdf
    27 Jan 2023: IHMM evaluation in (Beal et al.,2002) is more elaborate: it allows the IHMM to con-tinue learning about new data encountered during test-ing. ... Subjecting theseresults to the same analysis as the artificial data re-veals similar compared test-set
  16. 4F13 Machine Learning: Coursework #4: Reinforcement Learning Zoubin…

    https://mlg.eng.cam.ac.uk/teaching/4f13/1011/cw/coursework4.pdf
    19 Nov 2023: Test yourvalueIteration algorithm.
  17. G:\bioinformatics\Bioinfo-26(7)issue\btq053.dvi

    https://mlg.eng.cam.ac.uk/pub/pdf/LipGhaBor10.pdf
    13 Feb 2023: A two-sample test tries to decide whether twosamples, in our case x and DC , have been generated by the samedistribution or not. ... edge,and might artificially boost prediction accuracy if they appear in both trainingand test set.
  18. Relational Learning with Gaussian Processes Wei ChuCCLS Columbia…

    https://mlg.eng.cam.ac.uk/pub/pdf/ChuSinGhaetal07.pdf
    13 Feb 2023: in the input space and provides proba-bilistic induction over unseen test points. ... K̃(zm, zt)]T. One can computethe Bernoulli distribution over the test labelyt by.
  19. 4F13 Machine Learning: Coursework #1: Gaussian Processes Carl Edward…

    https://mlg.eng.cam.ac.uk/teaching/4f13/1415/cw/coursework1.pdf
    19 Nov 2023: Show and comment on the fit and the hypers, and the predictions forthe test data.
  20. Relational Learning with Gaussian Processes Wei ChuCCLS Columbia…

    https://mlg.eng.cam.ac.uk/zoubin/papers/relationalgp.pdf
    27 Jan 2023: in the input space and provides proba-bilistic induction over unseen test points. ... K̃(zm, zt)]T. One can computethe Bernoulli distribution over the test labelyt by.
  21. Gaussian Process Model Based Predictive Control

    https://mlg.eng.cam.ac.uk/pub/pdf/KocMurRasGir04.pdf
    13 Feb 2023: 4. Fitting of theresponse for validation signal:• average absolute test error. AE = 0.1276 (14). • ... average squared test error. SE = 0.0373 (15). 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100002.

Refine your results

Search history

Recently clicked results

Recently clicked results

Your click history is empty.

Recent searches

Recent searches

Your search history is empty.