Ensembles on Random Patches
نویسندگان
چکیده
Table 1: Accuracy on small datasets (in %). Validation RF ET P-DT P-ET RS-DT RS-ET RP-DT RP-ET diabetes 77.12 (6) 77.25 (5) 77.67 (4) 78.01 (3) 75.11 (8) 76.77 (7) 78.82 (2) 79.07 (1) dig44 94.99 (7) 95.78 (1) 91.86 (8) 95.46 (4) 95.07 (6) 95.69 (3) 95.13 (5) 95.72 (2) ionosphere 94.40 (6) 95.15 (3) 93.86 (8) 94.75 (5) 94.11 (7) 94.90 (4) 95.20 (2) 95.36 (1) pendigits 98.94 (7) 99.33 (1) 98.09 (8) 99.28 (4) 99.02 (6) 99.31 (3) 99.07 (5) 99.32 (2) letter 95.36 (7) 96.38 (1) 92.72 (8) 95.87 (4) 95.68 (6) 96.08 (3) 95.74 (5) 96.10 (2) liver 72.37 (5) 71.90 (6) 72.55 (4) 72.88 (3) 68.06 (8) 70.88 (7) 74.53 (1) 74.37 (2) musk2 97.18 (7) 97.73 (1) 96.89 (8) 97.60 (4) 97.58 (6) 97.72 (3) 97.60 (5) 97.73 (2) ring-norm 97.44 (6) 98.10 (5) 96.41 (8) 97.28 (7) 98.25 (4) 98.41 (3) 98.50 (2) 98.54 (1) satellite 90.97 (7) 91.56 (1) 90.01 (8) 91.40 (5) 91.31 (6) 91.50 (3) 91.41 (4) 91.54 (2) segment 97.46 (6) 98.17 (2) 96.78 (8) 98.10 (4) 97.33 (7) 98.14 (3) 97.52 (5) 98.21 (1) sonar 82.92 (7) 86.92 (3) 80.03 (8) 84.73 (5) 83.07 (6) 87.07 (2) 85.42 (4) 88.15 (1) spambase 94.80 (7) 95.36 (3) 93.69 (8) 95.01 (6) 95.01 (5) 95.50 (2) 95.11 (4) 95.57 (1) two-norm 97.54 (6) 97.77 (2) 97.52 (7) 97.59 (5) 97.46 (8) 97.63 (4) 97.76 (3) 97.82 (1) vehicle 88.67 (5) 88.68 (4) 88.26 (8) 88.74 (3) 88.41 (7) 88.60 (6) 89.22 (1) 89.21 (2) vowel 92.04 (5) 95.12 (1) 85.19 (8) 93.49 (4) 89.76 (7) 94.34 (3) 91.10 (6) 94.48 (2) waveform 85.45 (6) 85.96 (2) 84.89 (8) 85.68 (5) 84.91 (7) 85.69 (4) 85.85 (3) 86.21 (1) Avg. rank 6.25 2.5625 7.4375 4.4375 6.5 3.75 3.5626 1.5
منابع مشابه
Foveated analysis of image features at fixations
Analysis of the statistics of image features at observers' gaze can provide insights into the mechanisms of fixation selection in humans. Using a foveated analysis framework, in which image patches were analyzed at the resolution corresponding to their eccentricity from the prior fixation, we studied the statistics of four low-level local image features: luminance, RMS contrast, and bandpass ou...
متن کاملOn the similarity between Nakagami-m Fading distribution and the Gaussian ensembles of random matrix theory
We report the similarity between the Nakagami-m fading distribution and the three Gaussian ensembles of random matrix theory. We provide a brief review of random matrix theory and wireless fading. We show that the Nakagami-m distribution serves as mapping between the three ensembles. The statistics of the wireless fading amplitude, as modeled by Nakagami-m distribution, provide a rare example o...
متن کاملCluster Ensembles for High Dimensional Clustering: An Empirical Study
This paper studies cluster ensembles for high dimensional data clustering. We examine three different approaches to constructing cluster ensembles. To address high dimensionality, we focus on ensemble construction methods that build on two popular dimension reduction techniques, random projection and principal component analysis (PCA). We present evidence showing that ensembles generated by ran...
متن کاملThe Utility of Randomness in Decision Tree Ensembles
The use of randomness in constructing decision tree ensembles has drawn much attention in the machine learning community. In general, ensembles introduce randomness to generate diverse trees and in turn they enhance ensembles’ predictive accuracy. Examples of such ensembles are Bagging, Random Forests and Random Decision Tree. In the past, most of the random tree ensembles inject various kinds ...
متن کاملOblique Random Forests for 3-D Vessel Detection Using Steerable Filters and Orthogonal Subspace Filtering
We propose a machine learning-based framework using oblique random forests for 3-D vessel segmentation. Two different kinds of features are compared. One is based on orthogonal subspace filtering where we learn 3-D eigenspace filters from local image patches that return task optimal feature responses. The other uses a specific set of steerable filters that show, qualitatively, similarities to t...
متن کاملEntropy, Maximum Entropy Priciple and Quantum Statistical Information for Various Random Matrix Ensembles
The random matrix ensembles (RME) of quantum statistical Hamiltonians, e.g. Gaussian random matrix ensembles (GRME) and Ginibre random matrix ensembles (Ginibre RME), are applied in literature to following quantum statistical systems: molecular systems, nuclear systems, disordered materials, random Ising spin systems, and two-dimensional electron systems (Wigner-Dyson electrostatic analogy). Me...
متن کامل