Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Weighted KDE #26

Merged
merged 12 commits into from
Jun 27, 2017
36 changes: 21 additions & 15 deletions src/bivariate.jl
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ function default_bandwidth(data::Tuple{RealVector,RealVector})
end

# tabulate data for kde
function tabulate(data::Tuple{RealVector, RealVector}, midpoints::Tuple{Range, Range})
function tabulate(data::Tuple{RealVector, RealVector}, midpoints::Tuple{Range, Range}, weights::Weights = default_weights(data))
xdata, ydata = data
ndata = length(xdata)
length(ydata) == ndata || error("data vectors must be of same length")
Expand All @@ -41,17 +41,19 @@ function tabulate(data::Tuple{RealVector, RealVector}, midpoints::Tuple{Range, R

# Set up a grid for discretized data
grid = zeros(Float64, nx, ny)
ainc = 1.0 / (ndata*(sx*sy)^2)
ainc = 1.0 / (sum(weights)*(sx*sy)^2)

# weighted discretization (cf. Jones and Lotwick)
for (x, y) in zip(xdata,ydata)
for i in 1:length(xdata)
x = xdata[i]
y = ydata[i]
kx, ky = searchsortedfirst(xmid,x), searchsortedfirst(ymid,y)
jx, jy = kx-1, ky-1
if 1 <= jx <= nx-1 && 1 <= jy <= ny-1
grid[jx,jy] += (xmid[kx]-x)*(ymid[ky]-y)*ainc
grid[kx,jy] += (x-xmid[jx])*(ymid[ky]-y)*ainc
grid[jx,ky] += (xmid[kx]-x)*(y-ymid[jy])*ainc
grid[kx,ky] += (x-xmid[jx])*(y-ymid[jy])*ainc
grid[jx,jy] += (xmid[kx]-x)*(ymid[ky]-y)*ainc*weights[i]
grid[kx,jy] += (x-xmid[jx])*(ymid[ky]-y)*ainc*weights[i]
grid[jx,ky] += (xmid[kx]-x)*(y-ymid[jy])*ainc*weights[i]
grid[kx,ky] += (x-xmid[jx])*(y-ymid[jy])*ainc*weights[i]
end
end

Expand Down Expand Up @@ -87,41 +89,45 @@ end

const BivariateDistribution = Union{MultivariateDistribution,Tuple{UnivariateDistribution,UnivariateDistribution}}

function kde(data::Tuple{RealVector, RealVector}, midpoints::Tuple{Range, Range}, dist::BivariateDistribution)
k = tabulate(data,midpoints)
default_weights(data::Tuple{RealVector, RealVector}) = UniformWeights(length(data[1]))

function kde(data::Tuple{RealVector, RealVector}, weights::Weights, midpoints::Tuple{Range, Range}, dist::BivariateDistribution)
k = tabulate(data, midpoints, weights)
conv(k,dist)
end

function kde(data::Tuple{RealVector, RealVector}, dist::BivariateDistribution;
boundary::Tuple{Tuple{Real,Real}, Tuple{Real,Real}} = (kde_boundary(data[1],std(dist[1])),
kde_boundary(data[2],std(dist[2]))),
npoints::Tuple{Int,Int}=(256,256))
npoints::Tuple{Int,Int}=(256,256),
weights::Weights = default_weights(data))

xmid = kde_range(boundary[1],npoints[1])
ymid = kde_range(boundary[2],npoints[2])

kde(data,(xmid,ymid),dist)
kde(data,weights,(xmid,ymid),dist)
end

function kde(data::Tuple{RealVector, RealVector}, midpoints::Tuple{Range, Range};
bandwidth=default_bandwidth(data), kernel=Normal)
bandwidth=default_bandwidth(data), kernel=Normal, weights::Weights = default_weights(data))

dist = kernel_dist(kernel,bandwidth)
kde(data,midpoints,dist)
kde(data,weights,midpoints,dist)
end

function kde(data::Tuple{RealVector, RealVector};
bandwidth=default_bandwidth(data),
kernel=Normal,
boundary::Tuple{Tuple{Real,Real}, Tuple{Real,Real}} = (kde_boundary(data[1],bandwidth[1]),
kde_boundary(data[2],bandwidth[2])),
npoints::Tuple{Int,Int}=(256,256))
npoints::Tuple{Int,Int}=(256,256),
weights::Weights = default_weights(data))

dist = kernel_dist(kernel,bandwidth)
xmid = kde_range(boundary[1],npoints[1])
ymid = kde_range(boundary[2],npoints[2])

kde(data,(xmid,ymid),dist)
kde(data,weights,(xmid,ymid),dist)
end

# matrix data
Expand Down
51 changes: 33 additions & 18 deletions src/univariate.jl
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,10 @@ function default_bandwidth(data::RealVector, alpha::Float64 = 0.9)
return alpha * width * ndata^(-0.2)
end

function default_weights(data::RealVector)
UniformWeights(length(data))
end


# Roughly based on:
# B. W. Silverman (1982) "Algorithm AS 176: Kernel Density Estimation Using
Expand Down Expand Up @@ -66,23 +70,32 @@ function kde_range(boundary::Tuple{Real,Real}, npoints::Int)
lo:step:hi
end

immutable UniformWeights{N} end

UniformWeights(n) = UniformWeights{n}()

Base.sum(x::UniformWeights) = 1.
Base.getindex{N}(x::UniformWeights{N}, i) = 1/N

typealias Weights Union{UniformWeights, RealVector, WeightVec}


# tabulate data for kde
function tabulate(data::RealVector, midpoints::Range)
ndata = length(data)
function tabulate(data::RealVector, midpoints::Range, weights::Weights=default_weights(data))
npoints = length(midpoints)
s = step(midpoints)

# Set up a grid for discretized data
grid = zeros(Float64, npoints)
ainc = 1.0 / (ndata*s*s)
ainc = 1.0 / (sum(weights)*s*s)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

When you switch to WeightsVec, you can use its sum field to avoid recomputing its sum.


# weighted discretization (cf. Jones and Lotwick)
for x in data
for (i,x) in enumerate(data)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Even better: for (x, w) in zip(data, weights).

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I totally agree here, but this needs weights to implement the "Iterator protocoll", which is not given for weights::WeightVec or my (WIP) custom UniformWeights type.

k = searchsortedfirst(midpoints,x)
j = k-1
if 1 <= j <= npoints-1
grid[j] += (midpoints[k]-x)*ainc
grid[k] += (x-midpoints[j])*ainc
grid[j] += (midpoints[k]-x)*ainc*weights[i]
grid[k] += (x-midpoints[j])*ainc*weights[i]
end
end

Expand Down Expand Up @@ -119,30 +132,30 @@ function conv(k::UnivariateKDE, dist::UnivariateDistribution)
end

# main kde interface methods
function kde(data::RealVector, midpoints::Range, dist::UnivariateDistribution)
k = tabulate(data, midpoints)
function kde(data::RealVector, weights::Weights, midpoints::Range, dist::UnivariateDistribution)
k = tabulate(data, midpoints, weights)
conv(k,dist)
end

function kde(data::RealVector, dist::UnivariateDistribution;
boundary::Tuple{Real,Real}=kde_boundary(data,std(dist)), npoints::Int=2048)
boundary::Tuple{Real,Real}=kde_boundary(data,std(dist)), npoints::Int=2048, weights=default_weights(data))

midpoints = kde_range(boundary,npoints)
kde(data,midpoints,dist)
kde(data,weights,midpoints,dist)
end

function kde(data::RealVector, midpoints::Range;
bandwidth=default_bandwidth(data), kernel=Normal)
bandwidth=default_bandwidth(data), kernel=Normal, weights=default_weights(data))
bandwidth > 0.0 || error("Bandwidth must be positive")
dist = kernel_dist(kernel,bandwidth)
kde(data,midpoints,dist)
kde(data,weights,midpoints,dist)
end

function kde(data::RealVector; bandwidth=default_bandwidth(data), kernel=Normal,
npoints::Int=2048, boundary::Tuple{Real,Real}=kde_boundary(data,bandwidth))
npoints::Int=2048, boundary::Tuple{Real,Real}=kde_boundary(data,bandwidth), weights=default_weights(data))
bandwidth > 0.0 || error("Bandwidth must be positive")
dist = kernel_dist(kernel,bandwidth)
kde(data,dist;boundary=boundary,npoints=npoints)
kde(data,dist;boundary=boundary,npoints=npoints,weights=weights)
end

# Select bandwidth using least-squares cross validation, from:
Expand All @@ -152,10 +165,11 @@ end

function kde_lscv(data::RealVector, midpoints::Range;
kernel=Normal,
bandwidth_range::Tuple{Real,Real}=(h=default_bandwidth(data); (0.25*h,1.5*h)))
bandwidth_range::Tuple{Real,Real}=(h=default_bandwidth(data); (0.25*h,1.5*h)),
weights=default_weights(data))

ndata = length(data)
k = tabulate(data, midpoints)
k = tabulate(data, midpoints, weights)

# the ft here is K/ba*sqrt(2pi) * u(s), it is K times the Yl in Silverman's book
K = length(k.density)
Expand Down Expand Up @@ -194,8 +208,9 @@ function kde_lscv(data::RealVector;
boundary::Tuple{Real,Real}=kde_boundary(data,default_bandwidth(data)),
npoints::Int=2048,
kernel=Normal,
bandwidth_range::Tuple{Real,Real}=(h=default_bandwidth(data); (0.25*h,1.5*h)))
bandwidth_range::Tuple{Real,Real}=(h=default_bandwidth(data); (0.25*h,1.5*h)),
weights::Weights = default_weights(data))

midpoints = kde_range(boundary,npoints)
kde_lscv(data,midpoints; kernel=kernel, bandwidth_range=bandwidth_range)
kde_lscv(data,midpoints; kernel=kernel, bandwidth_range=bandwidth_range, weights=weights)
end
6 changes: 6 additions & 0 deletions test/bivariate.jl
Original file line number Diff line number Diff line change
Expand Up @@ -56,5 +56,11 @@ for X in ([0.0], [0.0,0.0], [0.0,0.5], [-0.5:0.1:0.5;])
@test all(k5.density .>= 0.0)
@test sum(k5.density)*step(k5.x)*step(k5.y) ≈ 1.0

k6 = kde([X X],(r,r);kernel=D, weights=ones(X)/length(X))
@test k4.density ≈ k6.density
end
end

k1 = kde([0.0 0.0; 1.0 1.0], (r,r), bandwidth=(1,1), weights=[0,1])
k2 = kde([1.0 1.0], (r,r), bandwidth=(1,1))
@test k1.density ≈ k2.density
6 changes: 6 additions & 0 deletions test/univariate.jl
Original file line number Diff line number Diff line change
Expand Up @@ -53,5 +53,11 @@ for X in ([0.0], [0.0,0.0], [0.0,0.5], [-0.5:0.1:0.5;])
@test all(k5.density .>= 0.0)
@test sum(k5.density)*step(k5.x) ≈ 1.0

k6 = kde(X,r;kernel=D, weights=ones(X)/length(X))
@test k4.density ≈ k6.density
end
end

k1 = kde([0.0, 1.], r, bandwidth=1, weights=[0,1])
k2 = kde([1.], r, bandwidth=1)
@test k1.density ≈ k2.density