### TABLE VI NETWORK DELAY AS A FUNCTION OF LOAD

### Table 2 shows that students of the same race are more likely to form a friendship than

2007

"... In PAGE 9: ...orrelation is always positive -- it ranges from .22 at Rice to .58 at Baylor. Segmentation of the Social Networks Table2 shows that the friendship networks at the 10 Texas universities are segmented by race, major, cohort, and political orientation. A variety of definitions and measures of 12 Newman (2003) and Jackson (2006) report cluster coefficients ranging from .... In PAGE 11: ...Table2 documents the absolute segmentation. If friendships were formed randomly, the distribution of characteristics among the friends of any subset of students should equal the distribution in the population.... In PAGE 11: ... In general, minorities tend to have more diverse social networks. Table2 also documents segmentation by major, cohort, and political orientation. Students have at least twice as many friends from the same major than random friend assignment would generate.... In PAGE 33: ...Table2 : Segmentation by Race, Major, Cohort and Political Orientation Rice U Texas Texas A amp;M SMU Baylor Texas Tech Texas Christian U North Texas UT Arlington Texas State Segmentation by Race Fraction of Students White/Hisp 0.82 0.... ..."

### Table 3. Predictions of metamodel for Cox proportional- hazards regression assuming five exposure groups, logarithmic within-worker standard deviation of 2, and a ratio of logarithmic between-worker variance to logarithmic group mean of 0.3; k = number of workers sampled on two random days per group True

2005

"... In PAGE 7: ...ious results (e.g. from a literature review) to guess the neighbourhood within which the true association parameter may lie. The predicted biases in Cox proportion-hazards regression for different plausible scenarios are illus- trated in Table3 . It is apparent that the metamodel predicts negligible attenuation when true risk is low, but the attenuation becomes more pronounced when the association parameter is large and the num- ber of sampled workers small.... ..."

### Table 5. Hierchical Logistic Regression Estimates of Interracial, Interethnic, and Intraethnic Friendships: Hispanic and Asian Ethnic Groups

"... In PAGE 26: ... In general, there is little evidence of panethnicity; what is most important to youth in selecting their best friends seems to be the search for co-ethnics. Table5 examines similar models, but here we focus on ethnic differences in the propensity to choose a same-ethnic, different-ethnic, or a different-race best friend. The outcome variables are similar to those of Table 5.... In PAGE 26: ... Table 5 examines similar models, but here we focus on ethnic differences in the propensity to choose a same-ethnic, different-ethnic, or a different-race best friend. The outcome variables are similar to those of Table5 . Models 1a, 2a, and 3a use Mexicans as the baseline ethnic category, while Models 1b, 2b, and 3b use Chinese as the baseline category.... In PAGE 27: ...composition are similar to those reported in Table 4, we only describe the variation among ethnic groups. [ Table5 about here.] Compared to Mexican youth, Chinese, Filipino, Japanese, and Vietnamese youth have lower odds of choosing a different-race friend over a same-ethnic friend.... In PAGE 29: ... Along with the fact that Koreans had relatively low odds of choosing a different-ethnic peer over a same-ethnic peer (see Models 2a and 2b), there is some evidence of low panethnic sentiment among Korean Americans. Figures 4 through 6 graphically display the predicted probability of selecting each type of friend by ethnicity, as predicted in the Models in Table5 . Here, we assume that youth attend a school which is 40% own race (about the average for Latinos, but above average for Asians), of whom 50% are co-ethnics (about the average for both Latinos and Asian American ethnic groups in our sample).... ..."

### Table 1. Average risk field extent S(m) (spans) and total risk exposure

2001

"... In PAGE 9: ...While Table1 gives the extent and absolute magnitude averages for each risk field, Figure 3 shows the results from a converse standpoint which is the distribution of risk exposure on non-maintenance spans in the network due to all possible single-span maintenance actions in its network. The data in Figure 3 pools individual trial Lm(i) values for the Modular H=5 designs over all networks for each maintenance type.... In PAGE 10: ... C. Discussion of results First, and not surprisingly, Table1 shows that in all trial networks the Type 3 maintenance model engenders much greater risk of restorability loss than with Type 2. Secondly, in comparing both risk field extent and total risk exposure as the design hop limit decreases from H = 6 down to H = 4 we see the expected contraction of extent and risk magnitude, due to changes in the design that increasingly keep restoration flows (and maintenance replacement paths) closer to home.... ..."

Cited by 2

### Table 2: Using tribes to predict risk.

"... In PAGE 5: ... Both training sets were constructed such that half of the reps were labeled high-risk and the other half were labeled low-risk . The scores of simple predictive rules using only the tribe atribute are shown in Table2 . These scores confirm that the tribes contain predictive information.... ..."

### Table 3. Predicted risk type for unknown HPVs

in Human Papillomavirus Risk Type Classification from Protein Sequences Using Support Vector Machines

"... In PAGE 8: ... If the text documents are unavailable for unknown HPVs, there is no way to clas- sify them, whereas the sequence-based classification does not need to use any additional information except sequence itself. Table3 shows the risk type prediction for HPVs marked as unknown in Ta- ble 1. HPV26, HPV54, HPV57, and HPV70 are predicted as high-, low-, low-, and high-risk, respectively.... ..."

### Table 6. Prediction of risk of positive laparotomy findings.

"... In PAGE 3: ...NS/LP vs. MC/LD, P = 0.04) (Table 5). Estimation of the probability ofsubdiaphragmatic disease Using the regression coefficients determined in the logis- tic regression analysis for the four independent predictive factors a mathematical model was derived to provide a quantitative estimate of the probability of subdiaphrag- matic disease for individual patients. All tested variants are summarized in Table6 . The observed and predicted probability for each variant is listed.... ..."

### Table 2. The manually classified risk types of HPVs. Type Risk Type Risk Type Risk Type Risk

"... In PAGE 2: ... Assuming that Table 2 below is correct, the risk type for four of 69 HPVs is not known, so that 65 HPVs are evaluated. Twenty among 65 HPVs are classified as high-risk and the remaining 45 are classified as low-risk, while there are only 12 high-risk HPVs in Table2 . Since 53 HPVs are correctly classified, the accuracy is 81.... In PAGE 4: ... Lastly, we used the comment of HPV types to classify some types difficult to be classified. Table2 shows the summarized classification of HPVs according to its risk. In the all experiments below, we used only lt;comment gt; part.... In PAGE 6: ... Table 5 shows the predicted risk type for the HPV types whose risks are not known exactly. These HPVs are described as null ?null in Table2 . According to previous research on HPV (Chan et al.... ..."