remove Mean/Variance

Author: joshday

docs/src/index.md

@@ -1,8 +1,3 @@
-```@raw html
-<img src="https://user-images.githubusercontent.com/8075494/32734476-260821d0-c860-11e7-8c91-49ba0b86397a.gif"
-    style = "display: block; margin-left: auto; margin-right: auto; width: 50%;">
-```
-
 # Home
 
 **OnlineStats** is a Julia package for statistical analysis with algorithms that run both **online** and **in parallel**.  Online algorithms are well suited for streaming data or when data is too large to hold in memory.  Observations are processed one at a time and all **algorithms use O(1) memory**.
@@ -94,4 +89,6 @@ fit!(g, itr)
 
 ```@raw html
 <iframe src="https://ghbtns.com/github-btn.html?user=joshday&repo=OnlineStats.jl&type=star&count=true&size=large" frameborder="0" scrolling="0" width="160px" height="30px"></iframe>
-```
+```
+
+![](https://user-images.githubusercontent.com/8075494/32734476-260821d0-c860-11e7-8c91-49ba0b86397a.gif)

src/OnlineStats.jl

@@ -3,7 +3,7 @@ module OnlineStats
 using RecipesBase, Reexport, Statistics, LinearAlgebra, Dates
 @reexport using OnlineStatsBase, LossFunctions, PenaltyFunctions, LearnBase
 
-import OnlineStatsBase: OnlineStat, name, _fit!, _merge!, eachrow, eachcol
+import OnlineStatsBase: OnlineStat, name, _fit!, _merge!, eachrow, Mean, Variance
 import LearnBase: fit!, nobs, value, predict, transform, transform!
 import StatsBase: autocov, autocor, confint, skewness, kurtosis, entropy, midpoints, 
     fweights, sample, coef, Histogram

src/stats/stats.jl

@@ -13,46 +13,6 @@ function print_stat_tree(io::IO, stats)
     end
 end
 
-#-----------------------------------------------------------------------# Variance
-"""
-    Variance(T = Float64; weight=EqualWeight())
-
-Univariate variance, tracked as type `T`.
-
-# Example
-
-    o = fit!(Variance(), randn(10^6))
-    mean(o)
-    var(o)
-    std(o)
-"""
-mutable struct Variance{T, W} <: OnlineStat{Number}
-    σ2::T
-    μ::T
-    weight::W
-    n::Int
-end
-function Variance(T::Type{<:Number} = Float64; weight = EqualWeight()) 
-    Variance(zero(T), zero(T), weight, 0)
-end
-Base.copy(o::Variance) = Variance(o.σ2, o.μ, o.weight, o.n)
-function _fit!(o::Variance{T}, x) where {T}
-    μ = o.μ
-    γ = T(o.weight(o.n += 1))
-    o.μ = smooth(o.μ, T(x), γ)
-    o.σ2 = smooth(o.σ2, (T(x) - o.μ) * (T(x) - μ), γ)
-end
-function _merge!(o::Variance, o2::Variance)
-    γ = o2.n / (o.n += o2.n)
-    δ = o2.μ - o.μ
-    o.σ2 = smooth(o.σ2, o2.σ2, γ) + δ ^ 2 * γ * (1.0 - γ)
-    o.μ = smooth(o.μ, o2.μ, γ)
-    o
-end
-value(o::Variance) = o.n > 1 ? o.σ2 * unbias(o) : 1.0
-Statistics.var(o::Variance) = value(o)
-Statistics.mean(o::Variance) = o.μ
-
 #-----------------------------------------------------------------------# AutoCov and Lag
 # Lag
 """
@@ -703,30 +663,6 @@ function _fit!(o::KMeans, x)
     end
 end
 
-#-----------------------------------------------------------------------# Mean
-"""
-    Mean(T = Float64; weight=EqualWeight())
-
-Track a univariate mean, stored as type `T`.
-
-# Example
-
-    @time fit!(Mean(), randn(10^6))
-"""
-mutable struct Mean{T,W} <: OnlineStat{Number}
-    μ::T
-    weight::W
-    n::Int
-end
-Mean(T::Type{<:Number} = Float64; weight = EqualWeight()) = Mean(zero(T), weight, 0)
-_fit!(o::Mean{T}, x) where {T} = (o.μ = smooth(o.μ, x, T(o.weight(o.n += 1))))
-function _merge!(o::Mean, o2::Mean)
-    o.n += o2.n
-    o.μ = smooth(o.μ, o2.μ, o2.n / o.n)
-end
-Statistics.mean(o::Mean) = o.μ
-Base.copy(o::Mean) = Mean(o.μ, o.weight, o.n)
-
 #-----------------------------------------------------------------------# Moments
 """
     Moments(; weight=EqualWeight())

test/runtests.jl

@@ -390,13 +390,6 @@ end
     o2 = fit!(LinReg(), OnlineStatsBase.eachrow(ymat[:,[4,1]], ymat[:,3]))
     @test coef(o, [.2,.4]; y=3, x = [4,1], bias=false) ≈ coef(o2, [.2, .4])
 end
-#-----------------------------------------------------------------------# Mean
-@testset "Mean" begin
-    o = fit!(Mean(), y)
-    @test value(o) ≈ mean(y)
-    @test mean(o) ≈ mean(y)
-    @test ≈(mergevals(Mean(), y, y2)...)
-end
 #-----------------------------------------------------------------------# ML
 @testset "ML" begin
     o = OnlineStats.preprocess(OnlineStatsBase.eachrow(ymat))
@@ -577,23 +570,6 @@ end
     @test ≈(mergevals(Sum(), y, y2)...)
     @test ==(mergevals(Sum(Int), z, z2)...)
 end
-#-----------------------------------------------------------------------# Variance
-@testset "Variance" begin
-    o = fit!(Variance(), y)
-    @test mean(o) ≈ mean(y)
-    @test var(o) ≈ var(y)
-    @test std(o) ≈ std(y)
-
-    @test ≈(mergevals(Variance(), x, x2)...)
-    @test ≈(mergevals(Variance(), y, y2)...)
-    @test ≈(mergevals(Variance(Float32), Float32.(y), Float32.(y2))...; atol=.001)
-    @test ≈(mergevals(Variance(), z, z2)...)
-    # Issue 116
-    @test std(Variance()) == 1
-    @test std(fit!(Variance(), 1)) == 1
-    @test std(fit!(Variance(), [1, 2])) == sqrt(.5)
-end
-
 
 include("test_kahan.jl")