Feature performance overview

Overview of Feature Values

The following measures were taken for the results of each feature:

obs (number of observations within the feature set)
NAs (number of NAs, i.e., missing values, within the feature set)
min (minimum), mean (arithmetic mean), max (maximum), std_dev (standard deviation)
1st_qu (1st quartile = lower quartile = 25%-quantile)
median (median = 50%-quantile)
3rd_qu (3rd quartile = upper quartile = 75%-quantile)
co_var (coefficient of variation = standard deviation / arithmetic mean)

	obs	min	qu_1st	med	mean	qu_3rd	max	sd	coeff_var
stacks	527	4	6	12	12	16	20	6	0.52
tiers	527	5	5	6	5.9	7	8	0.91	0.16
stack.tier.ratio	527	0.25	0.31	0.5	28	1.2	8.8e+02	1.5e+02	5.1
container.density	527	0.6	0.6	0.67	0.68	0.71	0.8	0.07	0.1
empty.stack.pct	527	0	0	0	10	0.1	3.8e+02	49	4.8
overstowing.stack.pct	527	0.5	0.8	1	84	1	8.8e+02	2.3e+02	2.8
group.same.min	527	0	0	0	0.49	1	1	0.5	1
group.same.max	527	1	1	4	3.7	6	14	2.9	0.79
group.same.mean	527	1	1	2.3	2.3	4.4	4.9	1.5	0.66
group.same.stdev	527	0	0	0.76	0.73	1.5	3.5	0.81	1.1
top.good.min	527	1	1	2	2	2	8	0.96	0.48
top.good.max	527	5	10	14	15	19	36	6.1	0.4
top.good.mean	527	3	5.7	7.2	2.6e+02	9.5	1.5e+04	1.7e+03	6.4
top.good.stdev	527	0.78	2.8	4.1	73	5.7	1.1e+04	6.5e+02	8.9
overstowage.pct	527	0.21	0.37	0.45	5	0.48	3.8e+02	38	7.5
bflb	527	8	18	29	30	41	72	14	0.46

Summary of feature steps

The following table summarizes the feature steps over all instances.

size (percentage of features affected by this step)
ok, ..., other (percentage of run status)
cost_min, cost_mean, cost_max (range of costs for this step)
cost_na (percentage of missing values occuring in costs)

	size	ok	timeout	memout	presolved	crash	other	unknown
all	16	1e+02	0	0	0	0	0	0

Duplicated Features

Underneath, you'll find blocks of duplicated features, i.e., observations that have equal values over all features. Please note, that only a maximum of 5 duplicates per block are shown.

There were no duplicated features.