get_quantile_scores(scores, quantiles = 0.0:0.01:1.0)

return the quantiles of the given N dimensional anomaly scores cube. quantiles (default: quantiles = 0.0:0.01:1.0) is a Float range of quantiles. Any score being greater or equal quantiles[i] and beeing smaller than quantiles[i+1] is assigned to the respective quantile quantiles[i].

Examples

julia> scores1 = rand(10, 2)
julia> quantile_scores1 = get_quantile_scores(scores1)

get_quantile_scores!{tp,N}(quantile_scores::AbstractArray{Float64, N}, scores::AbstractArray{tp,N}, quantiles::StepRangeLen{Float64} = 0.0:0.01:1.0)

return the quantiles of the given N dimensional scores array into a preallocated quantile_scores array, see get_quantile_scores().

compute_ensemble(m1_scores, m2_scores[, m3_scores, m4_scores]; ensemble = "mean")

compute the mean (ensemble = "mean"), minimum (ensemble = "min"), maximum (ensemble = "max") or median (ensemble = "median") of the given anomaly scores. Supports between 2 and 4 scores input arrays (m1_scores, ..., m4_scores). The scores of the different anomaly detection algorithms should be somehow comparable, e.g., by using get_quantile_scores() before.

Examples

julia> using MultivariateAnomalies
julia> scores1 = rand(10, 2)
julia> scores2 = rand(10, 2)
julia> quantile_scores1 = get_quantile_scores(scores1)
julia> quantile_scores2 = get_quantile_scores(scores2)
julia> compute_ensemble(quantile_scores1, quantile_scores2, ensemble = "max")