Skip to content

Resampling uncertain values

Uncertain values may be resampled by drawing random number from the distributions furnishing them.

Documentation

# UncertainData.Resampling.resampleMethod.

1
resample(uv::AbstractUncertainValue)

Sample the uncertain value once, drawing values from the entire support of the probability distribution furnishing it.

source

# UncertainData.Resampling.resampleMethod.

1
resample(uv::AbstractUncertainValue, n::Int)

Sample the uncertain value n times, drawing values from the entire support of the probability distribution furnishing it.

source

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

resample(uv_theoretical)
resample(uv_theoretical_fitted)
resample(uv_kde)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

n = 500
resample(uv_theoretical, n)
resample(uv_theoretical_fitted, n)
resample(uv_kde, n)

Resampling can also be performed with constraints.

  • resample(uv::AbstractUncertainValue, constraint::SamplingConstraint) samples the uncertain value once, drawing from a restricted range of the support of the the probability distribution furnishing it.
  • resample(uv::AbstractUncertainValue, constraint::SamplingConstraint, n::Int) samples the uncertain value n times, drawing values from a restricted range of the support of the the probability distribution furnishing it.

Available sampling constraints are:

  1. TruncateStd(nσ::Int)
  2. TruncateMinimum(min::Number)
  3. TruncateMaximum(max::Number)
  4. TruncateRange(min::Number, max::Number)
  5. TruncateLowerQuantile(lower_quantile::Float64)
  6. TruncateUpperQuantile(upper_quantile::Float64)
  7. TruncateQuantiles(lower_quantile::Float64, upper_quantile::Float64)

For full documentation of the constraints, see the available constraints in the menu.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value with the restriction that the sampled
# values must be higher than the 0.2-th quantile of the distribution
# furnishing the value.
resample(uv_theoretical, TruncateLowerQuantile(0.2))
resample(uv_theoretical_fitted, TruncateLowerQuantile(0.2))
resample(uv_kde, TruncateLowerQuantile(0.2))

n = 100
resample(uv_theoretical, TruncateLowerQuantile(0.2), n)
resample(uv_theoretical_fitted, TruncateLowerQuantile(0.2), n)
resample(uv_kde, TruncateLowerQuantile(0.2))
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value  with the restriction that the sampled
# values must be lower than the 0.95-th quantile of the distribution
# furnishing the value.
resample(uv_theoretical, TruncateUpperQuantile(0.95))
resample(uv_theoretical_fitted, TruncateUpperQuantile(0.95))
resample(uv_kde, TruncateUpperQuantile(0.95))

n = 100
resample(uv_theoretical, TruncateUpperQuantile(0.95), n)
resample(uv_theoretical_fitted, TruncateUpperQuantile(0.95), n)
resample(uv_kde, TruncateUpperQuantile(0.95))
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value with the restriction that the sampled
# values must be within the (0.025, 0.975) quantile range.
resample(uv_theoretical, TruncateQuantiles(0.025, 0.975))
resample(uv_theoretical_fitted, TruncateQuantiles(0.025, 0.975))
resample(uv_kde, TruncateQuantiles(0.025, 0.975))

n = 100
resample(uv_theoretical, TruncateQuantiles(0.025, 0.975), n)
resample(uv_theoretical_fitted, TruncateQuantiles(0.025, 0.975), n)
resample(uv_kde, TruncateQuantiles(0.025, 0.975))
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value with the restriction that the sampled
# values have -2 as a lower bound.
resample(uv_theoretical, TruncateMinimum(-2))
resample(uv_theoretical_fitted, TruncateMinimum(-2))
resample(uv_kde, TruncateMinimum(-2))

n = 100
resample(uv_theoretical, TruncateMinimum(-2), n)
resample(uv_theoretical_fitted, TruncateMinimum(-2), n)
resample(uv_kde, TruncateMinimum(-2))
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value with the restriction that the sampled
# values have 3 as an upper bound.
resample(uv_theoretical, TruncateMaximum(3))
resample(uv_theoretical_fitted, TruncateMaximum(3))
resample(uv_kde, TruncateMaximum(3))

n = 100
resample(uv_theoretical, TruncateMaximum(3), n)
resample(uv_theoretical_fitted, TruncateMaximum(3), n)
resample(uv_kde, TruncateMaximum(3))
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
using Distributions, UncertainData

# Generate some uncertain values
uv_theoretical = UncertainValue(Normal, 4, 0.2)
uv_theoretical_fitted = UncertainValue(Normal, rand(Normal(1, 0.2), 1000))
uv_kde = UncertainValue(rand(Gamma(4, 5), 1000))

# Resample the uncertain value with the restriction that the sampled
# values must have values on the interval [-1, 1]. We first sample once,
# then 50 times.
resample(uv_theoretical, TruncateRange(-1, 1))
resample(uv_theoretical_fitted, TruncateRange(-1, 1))
resample(uv_kde, TruncateRange(-1, 1))

n = 100
resample(uv_theoretical, TruncateRange(-1, 1), n)
resample(uv_theoretical_fitted, TruncateRange(-1, 1), n)
resample(uv_kde, TruncateRange(-1, 1))