"... In PAGE 5: ...arent are identical. Results appear in Table 1. We attempted to optimize the performance of each op- erator by tuning the following parameters: bias, pop- ulation size and number of trials. The results in able 2 are similar to those in Table1 , although PMX and cycle showed very little improvement despite the pa- rameter tuning. The three order crossover operators (Order #1, Order #2 and Position) have similar per- formance.... ..."

"... In PAGE 9: ... This procedure was repeated for 9 independently drawn training sets. The cumulative results of these 9 runs are given in Table4 (see [Ack88] for more details). CN2 apos;s original bias is the maximum generality bias.... In PAGE 9: ... An inductive system using this bias, having decided to create a disjunct that matches a particular set of training examples, selects a maximally general disjunct that matches those examples and no others. The de nitions produced by CN2 using this bias are described in Table 2 and the top row of Table4 . The problem of error-prone small disjuncts is evident in this data.... In PAGE 10: ... An inductive system using this bias, having decided to create a disjunct that matches a particular (small) set of training examples, selects the disjunct consisting of all the conditions that are satis ed by those examples. The middle row in Table4 describes the de nitions produced by CN2 using the maximum generality bias for large disjuncts and the maximum speci city bias for small disjuncts. In these de nitions, the large disjuncts are identical to those in the original de nitions, but the small disjuncts are maximally speci c instead of being maximally general.... In PAGE 11: ... Then a condition matching all the examples in S is added to G, according to this selective speci city bias, if and only if it matches fewer than 25% of the C2 examples in T ? S. The bottom row in Table4 describes the de nitions produced by CN2 using the maximum generality bias for large disjuncts and the selective speci city bias for small disjuncts. The small disjuncts produced using the selective speci city bias are superior to those produced using the other biases.... In PAGE 12: ...9%) 1583 3024 (52%) 10.2% Table 5: Same as Table4 , except here \small quot; means coverage 9 or less. ways.... ..."

"... In PAGE 4: ... In order to determine if this was an imbalance of the VOR rather than the addition of a fixed drift-bias, we quantified the VOR imbalance in darkness for the observer whose VOR is illustrated in Figure 1 over a range of sinusoidal frequencies and peak velocities. Table3 illustrates the VOR imbalance calculated from 60 sec of stimulation at the combinations of frequency and velocity ranges. The main effect was an increased imbalance of the VOR with increasing velocity of body rotation particularly at low temporal frequencies.... ..."

"... In PAGE 6: ... The accuracy values deteriorate also due to the high bias (see mean values espec. in Table9 ), while height accuracy also gets worse due to the suboptimal base/height ratio (see sensor elevation and azimuth in Table 1). Taking all above factors into account, it becomes clear that IKONOS has a very high geometric accuracy potential and with sophisticated matching algorithms a height accuracy of 0.... ..."

"... In PAGE 3: ... Existing intra city trips by bus in various cities of India are as shown in Table 2. Table2 shows a large variation in the share of bus trips even amongst cities of similar size. Clearly there are factors other than the population size of the city that are responsible for this large variation.... ..."