### Table III. Complexity of some programs and their extracted modules

1994

Cited by 6

### Table IV. Module complexity before and after extraction

1994

Cited by 6

### Table 1 presents the asymptotic complexity of our algorithms. The insertion and dispatch-query time of a method D1B4D8B5 are given also in terms of CUD1 BP CYBYD1CY. We found that CUD1 is typically small; in average across our data set it was BIBMBC.

"... In PAGE 6: ... Table1 : Time and space bounds of the new algorithms The implementation of subtyping in SI hierarchies is theoretically optimal, since the query- and update- time are constant in the worst case. (Similar results were reported by Dietz [18, 19].... In PAGE 7: ... (There are no upper bounds on the total runtime of the RD algorithm.) We also see in Table1 that the insertion time of a type D8 to the MI subtyping data structure depends also on CPD2CRCTD7D8D3D6D7B4D8B5. The amortized cost of inserting a method D1B4D8B5 depends on the size of the family BYD1 as well as the number of descendants of D8.... In PAGE 11: ... Moreover, we stress that by eliminating degenerate families we only made things more difficult for our incremental algorithm. As Table1 indicates, the time for inserting a message D1 depends (among other things) on D0D3CV CUD1. Clearly, the insertion time per message increases if we eliminate all cases in which CUD1 BP BD.... In PAGE 27: ... When the type is loaded, we traverse all such messages, and perform the algorithm described before. 7 Directions for Further Research Revisiting Table1 we would like that the time bounds for insertions in the MI-case would depend on the number of parents (children) rather than the number of ancestors (or descendants). The only sort of updates to the hierarchy which we allowed in this paper were the additions of types, along with their methods, at the bottom of the hierarchy.... ..."

### TABLE I THE SETUP FOR CALCULATING THE KOLMOGOROV COMPLEXITY OF A FITNESS FUNCTION. THE SEARCH SPACE IS ORDERED (IN SOME WAY), THE KC IS MEASURED FOR THE STRING DEFINED BY THE f(x) COLUMN

### Table 1. Time and Space Complexity of the HDDITM Fea- ture Extraction Modules

2001

"... In PAGE 3: ... 4.1 Computational Model The approximate time and space complexity of each of the four modules in the feature extraction process is shown in Table1 . The computational model incorporates the fol- lowing parameters: m is the average size of the content of the input URLs,... ..."

Cited by 3

### Table 1. Time and Space Complexity of the HDDITM Fea- ture Extraction Modules

"... In PAGE 3: ... 4.1 Computational Model The approximate time and space complexity of each of the four modules in the feature extraction process is shown in Table1 . The computational model incorporates the fol- lowing parameters: m is the average size of the content of the input URLs, L is the size of the Lexicon, LR is the number of Lexical rules, LRF is the size of the Lexical rules file, CR is the number of Contextual rules, CRF is the size of the Contextual rules file.... ..."

### Table 4. Complexity of extracted rules in terms of number of antecedents per rulebase

### Table 2. Kolmogorov entropy

"... In PAGE 21: ...6, but we also calculated the evolution of the entropy as the network learns. The results are shown in Table2 and clearly indicate that the process becomes less chaotic. For = 45, there is initially a slight decrease in the entropy, but then this stabilises, implying that not all of the chaotic behavi- our disappears.... ..."

### Table 6: Dynamic probabilistic inference: Estimated value of final state given first six observations. 500 repetitions.

2000

"... In PAGE 8: ... The problem was to infer the val- ue of the final state variable DCD8 given the observations DEBDBN DEBEBN BMBMBMBN DED8. Table6 again demonstrates that GIS has a sizeable advantage over standard importance sampling. (In fact, the greedy approach can be applied to particle filter- ing [IB96, KKR95] to obtain further improvements on this task, but space bounds preclude a detailed discussion.... ..."

Cited by 3