### Table 1. Correct, lost and swapped tracks of the aircraft for two tracking algorithms using the NN plot-to-track association logic and kinematics data.

2004

"... In PAGE 5: ... Two different filtering algorithms have been considered: an Extended Kalman Filter (EKF) and a Variable Structure - Interactive Multiple Model (VS-IMM) which uses kinematics constraints to improve the accuracy (for details, see [1]). Table1 and Table 2 show the results obtained respectively for the aircraft and the car in terms of number of correct, lost and swapped tracks. The results have been averaged over 1000 Monte Carlo trials.... ..."

Cited by 1

### Table 2. Correct, lost and swapped tracks of the car for two tracking algorithms using the NN plot-to-track association logic and kinematics data.

2004

"... In PAGE 5: ... Two different filtering algorithms have been considered: an Extended Kalman Filter (EKF) and a Variable Structure - Interactive Multiple Model (VS-IMM) which uses kinematics constraints to improve the accuracy (for details, see [1]). Table 1 and Table2 show the results obtained respectively for the aircraft and the car in terms of number of correct, lost and swapped tracks. The results have been averaged over 1000 Monte Carlo trials.... ..."

Cited by 1

### Table 5. Correct, lost and swapped tracks of the aircraft 1 for two tracking algorithms using the NN plot-to-track association logic and kinematics data.

2004

"... In PAGE 6: ... The scope of this simulation is to test the discrimination capability of the plot-to-track association algorithm in the worst case: two similar target moving in the same direction. The results concerning the NN plot-to-track logic are reported in Table5 and Table 6 (aircraft 1 is the manoeuvring one while aircraft 2 is the one going straight). The performance are worse than the ones of the case with the car because the measurement of the second aircraft is noisier due to the greater glint phenomenon.... ..."

Cited by 1

### Table 7. Correct, lost and swapped tracks of the aircraft 1 for two tracking algorithms using the plot-to-track association logic described in sections 2 and 3.

2004

"... In PAGE 6: ...5% 21.1% The results concerning the plot-to-track association logic using both kinematics data and the electromagnetic images are reported in Table7 and Table 8. Consider the EKF algorithm.... ..."

Cited by 1

### Table 2: Implementational Details of Standard and Proposed Tracking Algorithms

"... In PAGE 8: ... Both mechanisms employ a Kalman filter model whose observation and dynamic noise models are learnt directly from the data. The two methods are sum- marized in Table2 below. Data association is performed by searching predicted bound-... ..."

### Table 1. Association Between Common Subgroup Membership and The Realization of Interaction Between Actors

"... In PAGE 26: ... We compare time to complete group assignments for all three layouts . Table1 : Time to Complete Group Assignment Seeing Groups In Graph Layouts / McGrath, Blythe amp; Krackhardt Figure 1: Layout 2 Figure 1 : Layout 3 Table 1 shows that on average viewers took the least amount of time assigning groups for Layout 1.... In PAGE 26: ... We compare time to complete group assignments for all three layouts . Table 1: Time to Complete Group Assignment Seeing Groups In Graph Layouts / McGrath, Blythe amp; Krackhardt Figure 1: Layout 2 Figure 1 : Layout 3 Table1 shows that on average viewers took the least amount of time assigning groups for Layout 1.... In PAGE 57: ... Maximizing Frank apos;s criterion is equivalent to maximizing the odds ratio (AD/CB) of Table 1 . The odds ratio of Table1 is large to the extent that actors interact with members of their subgroups (cell D) and do not interact with members of other subgroups (cell A) . The odds ratio is small to the extent that actors do not interact with members of their subgroups (cell C) and actors interact with others who are not in their subgroup (cell B) .... In PAGE 57: ... Because the odds ratio is stochastic, with values on the diagonals essentially evaluated relative to the marginals, the odds ratio accommodates variation in the data, and thus allows the researcher to identify non- overlapping, but permeable, subgroup boundaries instead of overlapping subgroups of actors satisfying a fixed criterion (most of the criteria available in UCINET and STRUCTURE are not stochastic, and therefore generate overlapping subgroups --see Frank, 1993, Freeman, 1992, and Kadushin, 1995) . Given the stochastic criterion, Frank described a simple hill-climbing algorithm for identifying subgroups by iteratively reassigning actors so as to maximize the odds ratio of Table1 . Frank applied the algorithm to data indicating professional discussions (ranging from once a month [1] to daily [4]) among teachers in a high school called quot;Our Hamilton High quot; .... ..."

### Table 1. Algorithms and association matrices.

"... In PAGE 6: ... I is the identity matrix. The Algorithms We introduce seven algorithms ( Table1 ). One (Text) is a non-spreading activation algorithm used as a baseline for comparison.... In PAGE 6: ...Cocite CC = Cocitation Fused Citation FC = C+BC+(3*CC) Text none input; outputs Document x Document SAText SAT = Document x Document SATextFC SATFC = T+(3*FC) To leverage the citations and text data in RIV*, we created six spreading activation algorithms. The data used in each association matrix R is described in Table1 . Four methods use individual association matrices, i.... In PAGE 6: ... The remaining two methods, Fused Citation and SAText + Fused Citation (SATextFC), use a weighted combination of the other matrices to produce their final association matrices. The weightings appear in Table1 . The weights were selected manually1 to provide normalization across matrices.... ..."

### Table 7 - I-90 Project Web Page

2005

"... In PAGE 79: ... It is impossible to determine from this information if the phases that have been completed were within budget. Table 6 - I-5 Project Web Page EXPENDITURE PLAN Project Funding Expenditures prior to 7/1/2003 Remaining Funds Total Pre-existing State, Federal, and Other Partnership Funds $13,308,355 $3,126,204 $16,434,558 Total Available Funding $13,308,355 $3,126,204 $16,434,558 Amount Required to Complete Additional Project Stage(s) * $43,751,000 Estimated Total Project Cost $60,185,558 Financial data is current as of 08/06/2004 WSDOT PIN(s): 100536N, 100536P * No additional funding source identified Note: Program Item Numbers (PINs) are used by the Legislature to keep track of financial data associated with a project or segment of work Another example is the combined I-90 Project Total Project Cost Expenditure Plan, shown below as Table7 , and the June 2004 Quarterly Report Project Cost Summary, shown as Table 8. Both tables are taken from the same web page.... ..."

### Table 3: Mean length of reconstructed tracks, resolutions and pulls for primary tracks.

"... In PAGE 16: ... Ideally, the distributions of pulls should be unbiased and have a Gaussian core of unity. Table3 presents values of pulls and residuals for four parameters x, y, tx and ty of properly reconstructed tracks and the mean length of the tracks given in the number of associated hits for all three algorithms. As can be seen, pulls for the reconstructed tracks are typically wider than unity.... In PAGE 18: ...Table3 ) on hb-mu2 computer (Pentium III), the mean CPU time needed for CATS to reconstruct one mixed event was about 240 ms. A comparison of the computing time dependence on the number of superimposed inelastic events for CATS and RANGER is shown in Fig.... ..."

### Table 2. Participating sites in the Interactive Track

2005

Cited by 25