### TABLE 2. Example of truncation procedure used to determine smallest observable peak for each sample in a data set

2001

Cited by 1

### Table 2. Peak and Tail Probabilities

1991

"... In PAGE 4: ... This is what causes jobs to pile-up. Table2 shows the peak and tail probabilities for the values of b we studied. The significant peak for lower values of b is also... ..."

Cited by 1

### Table 6. LECS escape line data

1996

"... In PAGE 16: ... The probability that a uorescent photon will escape from the detection volume is geometry dependent, and has been measured for the LECS using data obtained at BESSY and PANTER. The values given in Table6 have been adopted for use in the LECS calibration and show that 1.68% of incident X-rays with energies just above the Xe LI edge produce uorescent X-rays which are lost from the instrument, re- sulting in escape peaks.... In PAGE 16: ... This gure is less than the Xe L shell uorescent yield of 8% due to the probability that an emitted uorescent photon is absorbed within the de- tector. Table6 gives the escape fractions for energies be- tween the Xe LIII and Xe LI edges. Above the Xe LI edge the escape fraction is assumed to decrease linearly with energy, falling to 1.... ..."

Cited by 2

### Table 2: Detection probabilities P (m; a; zs)

"... In PAGE 19: ...Table 2: Detection probabilities P (m; a; zs) The results of applying Eq. (9) for the three sets of assumptions and two values of zs are given in Table2 , and Fig. 7 shows the probability distribution over the m; a plane for the model corresponding to the last column.... In PAGE 20: ...5) 3 3 2 4 6 5 Uranus(5) 10 6 22 17 14 9 Uranus(10) 2 1 5 4 2 1 Jupiter 60 38 131 100 85 52 Saturn 6 3 18 12 8 3 Table 3: Number of planets detected. In the second of the three models in Table2 , where both the m=M and a=RE are invariant for each of the planets as the lens mass varies, the probabilities simply re ect the scaled values derived from Fig. 4.... In PAGE 20: ... Hence, the integral over the mass distribution (Eq. (9)) yields just this probability in Table2 . The detection probability is reduced somewhat for z = 0:8 in the table because the semimajor axis is the same as for z = 0:5 but the lensing zone has moved in with the Einstein ring radius leaving the Jupiters to the right of the peak in Fig.... ..."

### Table 5: Peak detection: percentage of missed and extra peaks

"... In PAGE 48: ... The most important information of each table was extracted and collected into Tables 5 to 8. The percentage of missed plus extra peaks per the number of sinusoids in the original sig- nal is presented in Table5 . Algorithm set 7, which uses F-test in peak detection has clearly more errors than the others in some sections.... ..."

### Table 1: Peaks and troughs for US unemployment based on non-smoothed and smoothed conditional probabilities.

"... In PAGE 16: ... A trough corre- sponds with the last observation in a recession and a peak with the last observation in an expansion. The rst two columns of Table1 display the peaks and troughs resulting from the conditional probabilities. The peaks and troughs correspond reasonably well with the o cial NBER peaks and troughs displayed in the last two columns of Table 1 except for the recession in the 1990s.... In PAGE 16: ... The smoothed conditional probability that t 6 = 0 at time t is the average of the conditional probabilities at time t ? 1, t and t + 1. The third and fourth columns of Table1 show the peaks and troughs based on these smoothed conditional probabilities. Now we also detect the recession in the beginning of the 1990s.... In PAGE 17: ...alue is about 1.037) and is only temporary such that shocks eventually will die out. The magnitudes of the autoregressive parameter ( + t) show that the recession in the beginning of the 1990s was less severe than the other recessions. To show the in uence of shocks during recessions on future values of the unemployment rate, we display the impact of a unit shock in the rst month of each recession (based on the turning point results in Table1 ) in Figure 5. Hence we impose a unit shock in 1969.... ..."

### Table 11: Peak detection: percentage of missed and extra peaks

### Table 1. Photon and total error for a single observation (10 elementary exposures).

755

"... In PAGE 2: ... The measurement error expected for individual observations depends on the instrumental parameters, here taken as in Lindegren amp; Perryman (1996), on the magnitude of the star and on the quality of the metrology control; the results of Casertano et al. (1996) are reported in Table1 for reference. The new element added here is that the instanta- neous `true apos; position of each star includes the gravi- tational perturbation (Keplerian motion) induced by a single, non-luminous planetary mass orbiting the star.... In PAGE 2: ... If ap is in AU, and D in parsec, then is expressed in arc- seconds. We generally assume a single-observation measurement error = 10 as, appropriate to a star brighter than V 12 mag (see Table1 ), corre- sponding to the Sun at 200 pc. We also carried out tests with di erent measurement errors, and demonstrated explicitly that the detec- tion probability depends exclusively on the `signal- to-noise apos; ratio: S=N = = so that rescaling to di erent measurement errors is straightforward.... In PAGE 4: ... Conver- gence of the t, and not the signi cance of the values derived for a, is taken as the indicator of a positive detection.1 Observations for Earth-Sun systems were simulated with an error = 1 as as indicated in Table1 for nearby solar stars, in the case of a metrol- ogy accuracy of 20 pm. As for the case of the 2 test, the same simulations were repeated without as- trometric signatures in order to assess the probability of false detections.... ..."

### Table 1. Detected peaks for the different calibration sources.

"... In PAGE 4: ...1. SPECTRAL CALIBRATION Table1 shows the measured peaks for the different calibration sources. In figure 2 the result of the linear regression for all combinations is plotted.... ..."

### Table 3. Results with peak alignment and detection

2005

"... In PAGE 4: ...ndicated in Section 3.2.2. As our learning sets contain two or four replicas for each patient, we have removed all data corresponding to a single patient in each leave-one-out round, so as not to bias the estimates. Table3 reports the results obtained with pre-processing by peak alignment and detection. All ensemble methods are quite close but single trees are clearly inferior.... In PAGE 7: ... Table 7 shows the results obtained in this way for increasing values of M (R = 4). We observe that this very simple approach indeed improves with respect to the results of Table3 : for M = 2, both sensitivity and specificity increase (and error rate decreases from 10.44 to 7.... ..."