### Table Size for Quadratic Interpolation

1998

Cited by 1

### Table XIV. Formula sizes of the quadratic step semantics encodings vs. the BLACKBOX encoding

2005

Cited by 5

### Table 3: Execution times (minutes) for creating 128 clusters on collections varying in size. Ideal performance is a quadratic increase as the collection size increases and linear decrease as the number of nodes increases

### Table 2: For data from quadratic model, table shows the minimum region sizes required by

1998

Cited by 5

### Table 1: Absolute times for GA simulation on different processors. Problem: Quadratic Assignment Problem, Problem Size = 30 Population Size = 120, Number of Generations = 250

1995

Cited by 1

### Table 2: For data from quadratic model, table shows the minimum region sizes required by model selection criteria at different values of to perform within 3% of their maximum suc- cess rate.

1998

Cited by 5

### Table 4 shows the CPU time needed for the baseline implementation. Note that the CPU time is domi- nated by the effort for word moving, which is quadratic in G, not by the corpus size. For 500 word classes, there is a speedup by a factor of four and more as compared to the corresponding CPU times of the refined implementation given in Table 3.

1998

Cited by 25

### Table 1. Class groups of some imaginary quadratic orders

1999

"... In PAGE 7: ... We present the results of applying our algorithm to compute the class groups of four imaginary quadratic orders with various size discriminants. Table1 gives the discriminant, factorization of the discriminant (computed from a system of generators of the class group), class number, and elementary divisors of the class group for each of these orders. The number in parenthesis after the discriminant is the number of decimal digits.... ..."

Cited by 8

### Table 2: Sizes of norm minimization problems, SDP relaxation of quadratic programs with box constraints, SDP relaxation of the maximum cut problems, and SDP relaxation of graph partition problems.

2004

"... In PAGE 12: ...e., norm minimization problems, SDP relaxation of quadratic programs with box constraints, SDP relaxation of max-cut problems over lattice graphs, and SDP relaxation of graph partitioning problems, are shown at Table2 . The original formulation of graph partitioning problems gives a dense aggregate sparsity pattern not allowing us to use the conversion method, and therefore we previously applied an appropriate congruent transformation [8, Section 6] to them.... ..."

### Table 3: Influence on training time when varying the size of the quadratic program Q in SVMlight, when using the linadd formulation of the WD kernel. While training times do not vary dramatically one still observes the tendency that with larger sample size a larger Q becomes optimal. The Q = 112 column displays the same result as column LinWD1 in Table 1.

2006

"... In PAGE 23: ...e speed up by a factor of Q (i.e. the size of the quadratic subproblems, termed qpsize in SVMlight). However, there is a trade-off in choosing Q as solving larger quadratic subproblems is expensive (quadratic to cubic effort). Table3 shows the dependence of the computing time from Q and N. For example the gain in speed between choosing Q = 12 and Q = 42 for 1 million of examples is 54%.... ..."

Cited by 24