### Table 1: End-client and Network Heterogeneity

"... In PAGE 1: ... Though multicast applications reap enormous performance bene#0Cts from the underlying multicast service, they are fun- damentally challenged by the heterogeneity that is inher- ent in the disparate technologies that comprise the Internet, both within the end systems and across the network infras- tructure. Table1 shows the high variance in client and network capabilities today. End devices range from sim- ple palm-top personal digital assistants #28PDAs#29 to power- ful high-end desktop PCs, while network link characteristics can vary by many orders of magnitude in terms of delay, capacity, and error rate.... ..."

### Table 2: Sources of sampling uncertainty in sampling theory* Source Description Fundamental sampling error (FSE) A result of the constitutional heterogeneity (the particles being chemically of

2006

"... In PAGE 16: ... A modelling approach might use these effects as the basis for a mathematical model. In addition, sampling theory identifies eight distinct sources of error in sampling ( Table2 ); each of these can also be reduced to a variety of causal factors, which in turn can be used in various modelling approaches. An alternative approach is to consider all of the steps in the measurement process (Figure 1) as sources of uncertainty that make some contribution to the uncertainty of the final measurement.... In PAGE 16: ... As long as this requirement is met, any categorisation scheme may be applied to the estimation of uncertainty. The categorisation schemes listed in Table2 and Table 3 cover all practically important effects. 5.... In PAGE 27: ... Otherwise sample or increment extraction error (IXE) is created. Sample preparation errors (IPE) have several potential causes listed in Table2 , two of which are excluded as gross errors by the GUM definition. 10.... ..."

### Table 1: Capabilities and heterogeneous service support of WISCS. basic bit rate

1996

"... In PAGE 2: ...igure 1: Mobility vs. bit rate. High data rate applications and mobile low rate applications exhibit a great amount of heterogeneity in terms of bit rate, mobility, delay, tolerable bit error rate (BER), coverage, burstiness, and by the consumer acceptable equipment cost. Table1 lists the capabilities and heterogeneous service support of WISCS. Table 1: Capabilities and heterogeneous service support of WISCS.... ..."

Cited by 1

### Table 3: Heterogeneity, VAF and replication values of TRN

"... In PAGE 14: ...RN model starts with a set of unordered units (i.e., there is no predetermined grid of units) and the adjacency structure between prototypes is constructed dynamically during the adaptation process. - - - Table3 - - - Therefore, the analyst can restrict his investigations for model selection to the MSE (quantization error) and VAF measures. As indicated in table 3 for a decreasing sequence of number of prototypes, MSE slightly levels o when con- structing a partition on the basis of eight instead of nine TRN prototypes (also... ..."

### Table 3: Convergence rates for heterogeneous material a = 100.

1999

"... In PAGE 3: ... h0 corresponds to the maximum of the diameters of the elements in the coarse mesh. For a heterogeneous material, the convergence rates in the H1 0 norm are shown in Table 2, Table3 . The plot of the error versus the number of degrees of freedom is shown on Fig.... ..."

Cited by 2

### Table 2: Heterogeneity, similarity, VAF and replication values of SOMs

"... In PAGE 12: ...uitability of SOM formats are determined heuristically, e.g. by examining quot;goodness of t quot; measures against a sequence of shrinking SOM layers. - - - Table2 - - - Two such measures are listed in table 2: As a heterogeneity measure, the quot;Mean Squared Errors quot; (MSE) inform about relative distances between data points and corresponding prototype vectors wij. Since more and more dis- similar preference patterns are resembled together, total heterogeneity of cluster solutions should increase with decreasing numbers of prototypes.... In PAGE 13: ... However, the topological quality of those maps is achieved at the cost of the centroid property of prototype vectors. Table2 provides evidence for this trade-o relationship by comparing the quot;Variance-Accounted-For quot; (VAF) statistics of the partitions (within groups divided by total variance) with a quot;Corrected VAF quot; (CVAF) measure ad- justed for deviations of prototypes from respective class means (which again is inversely related to MSE): The spread between these two measures of total data recovery increases with improved topological quality of the map. The issue of cluster validity is addressed in two ways here: First, as a measure of partition agreement the Hubert and Arabie [16] adjusted Rand-Index ( apos;A.... ..."

### TABLE 4: Effects of Prizes and Heterogeneity on Difference of Games Won Between the Favorite and Underdog Per Match

2008

### Table 1. Average values and std. error of Hyper-volume values computed for Heterogeneous H-MOPSO (Hetero.H-MOPSO), C-MOPSO and Homogenous H-Mopso (Homo. H-MOPSO) methods after every cycle.

"... In PAGE 7: ... H-MOPSO, on the other hand, can run more subswarms on the faster processors and thus can make better use of the available processing power in a hetero- geneous environment. Evaluations Table1 shows the average hyper-volume measures obtained for both methods in heterogeneous and homogenous environments (for C-MOPSO, heteroge- neous and homogeneous are identical as explained above). Comparing C-MOPSO and H-MOPSO on the homogeneous environment, C-MOPSO performs better.... ..."

### Table 4: Sample averaged ML parameter estimates for the yogurt, catsup and detergent categories for the error-correction duration model, with household heterogeneity. Signiflcant estimates (at 5-% level) are given in boldface, and in parentheses we indicate the segments for which the segment-level estimate is signiflcant.

"... In PAGE 18: ... Furthermore, as the common-factor model is rejected against the error-correction model, short-run efiects apparently difier from the corresponding long-run efiects. In Table4 and Table 5, we present the estimation results for the three categories for the flnal models. Table 4 shows the mean parameters over the sample, that is PM m=1 ^ pm^ m.... In PAGE 18: ... In Table 4 and Table 5, we present the estimation results for the three categories for the flnal models. Table4 shows the mean parameters over the sample, that is PM m=1 ^ pm^ m. Parameters in boldface are signiflcant at 5%.... In PAGE 20: ... For display, it therefore holds that the long-run efiect of a permanent display is smaller than the direct efiect. Actually the long-run efiect for display is not signiflcantly difierent from zero for any segment, see Table4 . Display therefore mainly has short-run efiects, while price has short as well as long-run efiects.... ..."

### Table 3 Heterogeneous causal e ect estimates Bootstrapped:

"... In PAGE 15: ... Correlations between unobserved determinants of drinking and em- ployment vary across drinking status, with much higher positive correlation for non-drinkers than for drinkers. Table3 presents estimates of the mean ATE, TT, LATE, and MTE, the standard deviation of these measures across individuals at the maximum likelihood structural estimates, resampled standard errors of the mean esti- mates, and the 5th and 95th percentiles of the bootstrap distributions of the mean e ects. The LATE is de ned with respect to a policy change: The tax on beer is varied from its sample minimum (0.... ..."