Evaluation Framework 143 MSIA 6.8 4.8 2.8 1.8 R ou tin g O ve rh ea d (% ) 0 0.5 1 1.5 2 0 0.5 1 1.5 2 DGA OSPF AntNet BeeHive (a) Routing overhead MSIA 6.8 4.8 2.8 1.8 S ub op tim al R ou tin g O ve he ad [...] shortest hop count path”. Formally we could define the parameter as So = ∑n d=1 ∑n s=1 ∑k i=1(h sd i − hsd o ) × Lsd i Bt , s = d (1) A Performance Evaluation Framework 141 where n is total number of nodes [...] AntNet 24.9 99.8 97.1 75 327 2674 2938 BeeHive 25.3 99.9 99.4 28 86 2627 2783 2.8 DGA 32.0 73.6 30.0 1449 3259 2891 3623 OSPF 36.1 83.2 49.9 706 1930 2680 2987 AntNet 43.5 99.2 90.1 210 699 2797 3256 BeeHive …

to varying time allowance for outlier detection accuracy 0 8 1 accuracy 0,6 0,8 0 2 0,4 0 0,2 0 0 0 5 1 0 1 5 2 0 2 5 3 0 3 5 time0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 time (c) Prototypical anytime algorithm: [...] is given by the following equation: g(o i , e s ) = 1 (2⇡)d/2 · det(⌃ es)1/2 e ( 1 2 (oiµes ) T⌃1 es (oiµes )) where det(⌃ es) is the determinant and ⌃1 es the inverse of ⌃ es . The important part is that [...] solution to the problem. 3.1 Problem definition Outlier detection is the data mining task concerned with identifying deviating, abnormal data. As illustrated in Fig- ure 1(a), the idea is to compare objects …