### Table 2: MSE results of real images

"... In PAGE 8: ... These images are down-sampled, magnified by different methods by a factor of two and com- pared to the original images. We perform the MSE compar- ison on R,G,B channels independently and Table2 reported the averaged MSE values / standard deviations over 20 im- ages from the test set. BC is the bicubic interpolation, EXT is our extended method, BL is bilinear interpolation, DDT is our basic method and PR is pixel replication.... ..."

### Table 5.7: Nonlinear versus Linear Posterior Distributions Real Data Nonlinear (SMC filter)

2005

Cited by 6

### Table VII. Nonlinear versus linear posterior distributions real data

2005

Cited by 6

### Table VIII. Nonlinear versus linear moments real data

2005

Cited by 6

### Table 2: Structural Change and Nonlinearity in Mean and Volatility: Real- Time Seasonally Adjusted Industrial Production Structural Change Nonlinearity

in Are Statistical Reporting Agencies Getting It Right? Data Rationality and Business Cycle Asymmetry

### Table 5.2. The total costs for the real nonlinear system

### Table 5.2. The total costs for the real nonlinear system

### Table 2. Real cubic elds: Computation h+ and theory H +

"... In PAGE 4: ...e use (2.1) to compute f(x) and thereby deduce h+(x)=g+(x)+(1=3)f(x). Our hope that the error term in Conjecture 3.1 can be reduced to o(x1=2)would require that the numbers in the last column of Table 1 and the next to last column of Table2 tend to zero. The data seems rather borderline in support of o(x1=2), but strongly indicative that our conjecture as stated is true, and moreover that the o(x5=6) can be reduced at least somewhat.... In PAGE 4: ...3), at most one linear combination g+(x)+uf(x)can agree with H +(x)towithino(x1=2). The last two columns of Table2 suggest that u =1=3 is more likely than u =0. Infactu =0:31 ::: would give the best t, in a least squares sense, and this agrees well with the theoretically favored u =1=3.... ..."

Cited by 3

### Table 1: Linear OFRQP, linear MPSVM and nonlinear MPSVM, training and tenfold testing correctness and running times. Exe- cution times include tenfold training. Best results for the linear case (first two columns) are in bold.

2001

"... In PAGE 2: ... Section 5 contains numerical test results on six public data sets for both the linear and nonlinear MPSVM. These tests show a speedup of as high as 2169-times for our linear MPSVM over conventional SVM with nearly equal test set correctness ( Table1 , Vehicle Dataset). These tests also show that a nonlinear MPSVM can improve tenfold test set correctness over a conventional linear SVM from 59.... In PAGE 2: ...3% to 98.9% ( Table1 , Vowel Dataset). Simple and short MATLAB [18] codes, very similar to those of PSVM [12] underly the proposed MPSVM algorithms.... ..."

Cited by 9