statsample-glm 0.0.1 → 0.1.0

data/lib/statsample-glm/glm/mle/probit.rb

@@ -0,0 +1,100 @@
+require 'statsample-glm/glm/mle/base'
+
+module Statsample
+  module GLM
+    module MLE
+      # Probit MLE estimation.
+      # == Usage:
+      # 
+      #   mle = Statsample::MLE::Probit.new
+      #   mle.newton_raphson(x,y)
+      #   beta       = mle.parameters
+      #   likelihood = mle.likelihood(x,y,beta)
+      #   iterations = mle.iterations
+      class Probit < Statsample::GLM::MLE::Base
+
+       protected
+        def measurement data_set, coefficients
+          (data_set * coefficients).map { |x| Distribution::Normal.cdf(x) }
+        end
+        # F(B'Xi)
+        if  Statsample.has_gsl?
+          # F(B'Xi)
+          def f(b,x)
+            p_bx=(x*b)[0,0] 
+            GSL::Cdf::ugaussian_P(p_bx)
+          end
+          # f(B'Xi)
+          def ff(b,x)
+            p_bx=(x*b)[0,0] 
+            GSL::Ran::ugaussian_pdf(p_bx)
+          end
+        else
+          def f(b,x) #:nodoc:
+            p_bx=(x*b)[0,0] 
+            Distribution::Normal.cdf(p_bx)
+          end
+          def ff(b,x) #:nodoc:
+            p_bx=(x*b)[0,0] 
+            Distribution::Normal.pdf(p_bx)
+          end
+        end
+        # Log Likehood for x_i vector, y_i scalar and b parameters
+        def log_likelihood_i(xi,yi,b)
+          fbx=f(b,xi)
+          (yi.to_f*Math::log(fbx))+((1.0-yi.to_f)*Math::log(1.0-fbx))
+        end
+        # First derivative of log-likelihood probit function
+        # x: Matrix (NxM)
+        # y: Matrix (Nx1)
+        # p: Matrix (Mx1)
+        def first_derivative(x,y,b)
+          raise "x.rows!=y.rows" if x.row_size!=y.row_size
+          raise "x.columns!=p.rows" if x.column_size!=b.row_size            
+          n = x.row_size
+          k = x.column_size
+          fd = Array.new(k)
+          k.times {|i| fd[i] = [0.0]}
+          n.times do |i|
+            xi = Matrix.rows([x.row(i).to_a])
+            fbx=f(b,xi)
+            value1 = (y[i,0]-fbx)/ ( fbx*(1-fbx))*ff(b,xi) 
+            k.times do |j|
+              fd[j][0] += value1*xi[0,j]
+            end
+          end
+          Matrix.rows(fd, true)
+        end
+        # Second derivative of log-likelihood probit function
+        # x: Matrix (NxM)
+        # y: Matrix (Nx1)
+        # p: Matrix (Mx1)
+        
+        def second_derivative(x,y,b)
+          raise "x.rows!=y.rows" if x.row_size!=y.row_size
+          raise "x.columns!=p.rows" if x.column_size!=b.row_size
+          n = x.row_size
+          k = x.column_size
+          if Statsample.has_gsl?
+            sum=GSL::Matrix.zeros(k)
+          else
+            sum=Matrix.zero(k)
+          end
+          n.times do |i|
+            xi=Matrix.rows([x.row(i).to_a])
+            fbx=f(b,xi)
+            val=((ff(b,xi)**2) / (fbx*(1.0-fbx)))*xi.t*xi
+            if Statsample.has_gsl?
+              val=val.to_gsl
+            end
+            sum-=val
+          end
+          if Statsample.has_gsl?
+            sum=sum.to_matrix
+          end
+          sum
+        end
+      end # Probit
+    end # MLE
+  end # GLM
+end # Statsample

data/lib/statsample-glm/glm/normal.rb

@@ -0,0 +1,17 @@
+require 'statsample-glm/glm/base'
+
+module Statsample
+  module GLM
+
+    class Normal < Statsample::GLM::Base
+
+      def initialize data_set, dependent, opts={}
+        super data_set, dependent, opts
+      end
+
+      def to_s
+        "Statsample::GLM::Normal"
+      end
+    end
+  end
+end

data/lib/statsample-glm/glm/poisson.rb

@@ -0,0 +1,17 @@
+require 'statsample-glm/glm/base'
+
+module Statsample
+  module GLM
+
+    class Poisson < Statsample::GLM::Base
+
+      def initialize data_set, dependent, opts={}
+        super data_set, dependent, opts
+      end
+
+      def to_s
+        "Statsample::GLM::Poisson"
+      end
+    end
+  end
+end

data/lib/statsample-glm/glm/probit.rb

@@ -0,0 +1,16 @@
+require 'statsample-glm/glm/base'
+
+module Statsample
+  module GLM
+    class Probit < Statsample::GLM::Base
+      
+      def initialize data_set, dependent, opts={}
+        super data_set, dependent, opts
+      end
+
+      def to_s
+        "Statsample::GLM::Probit"
+      end
+    end
+  end
+end

data/lib/statsample-glm/version.rb

@@ -0,0 +1,5 @@
+module Statsample
+  module GLM
+    VERSION = "0.1.0"
+  end
+end

data/spec/data/logistic.csv

@@ -0,0 +1,51 @@
+x1,x2,y
+0.537322309644812,-0.866655707911859,0
+-0.717124209978434,-0.367820249977585,0
+-0.519166718891331,0.361486610435,1
+0.434970973986765,0.857332626245179,0
+-0.761822002215759,0.133438466268095,1
+1.51170030921189,0.716104533073575,1
+0.883854199811195,1.77206093023382,1
+-0.908689798854196,-0.10136697295802,1
+1.70331977539793,-0.777086491435508,0
+-0.246971150634099,-0.204573554913706,1
+-1.59077593922623,0.963353531412233,1
+-0.721548040910253,-1.10103024900542,1
+0.467025703920194,-0.404372761837392,1
+-0.510132788447137,-0.230226345183469,0
+0.430106510266798,0.0363730246866971,1
+-0.144353683251536,-0.838265540390497,0
+-1.54943800728303,1.12543549657924,1
+0.849307651309298,-0.57929175648001,1
+-0.640304240933579,-0.747060244805248,0
+1.31462478279425,0.58946979365152,1
+-0.399783455165345,-0.531952663697324,0
+0.0453055645017902,1.53338594419818,1
+-2.58212161987746,0.521992029051441,1
+-1.16484414309359,1.41631763288724,1
+-1.08829266466281,0.611402316795129,1
+-0.243893919684792,-0.518355638373296,0
+-1.96655661929441,-0.515192557101107,0
+0.301335373291024,-0.672697937866108,1
+-0.665832694463588,1.84347042325327,1
+-0.0120650855753837,-0.21195540664804,0
+1.5116066367604,-0.269869371631611,0
+0.557300353673344,0.296155694010096,1
+1.12829931872045,-2.18097898069634,0
+0.234443748015922,-1.21314663927206,0
+-2.03486690662651,1.49193669881581,1
+0.275544751380246,1.38969280369493,1
+-0.231465849558696,-0.400680808117106,0
+-0.356880153225012,-1.87282814976479,0
+-0.57746647541923,1.82394870451051,1
+1.35758352580655,0.637864732838274,1
+1.23971669378224,-0.141155946382493,0
+-0.662466275100489,0.0699950644281617,1
+0.313263561921793,1.32568550595165,1
+-1.08783223256362,-0.412599258349398,1
+1.41964722846899,0.14436832227506,1
+1.29325100940785,-1.16507785388489,0
+0.72153880625103,-2.16782049922428,0
+0.440580131022748,0.24318371493798,0
+0.0351917814720056,0.258954871320764,1
+-0.142353224879252,-0.151966534521183,1

data/spec/data/logistic_mle.csv

@@ -0,0 +1,201 @@
+a,b,c,y
+0.751712137198122,-3.26835910896471,1.70092606756102,0.0
+0.554214068711979,-2.95659724374009,2.66368360383625,0.0
+-1.85331644460541,-2.8293733798982,3.34679611669776,0.0
+-2.88610159022917,-0.738982407138072,4.74970154250585,0.0
+-2.60553098049408,0.561020310042032,5.48308397712906,0.0
+-4.27353214430849,1.62383436799538,5.35813425193809,0.0
+-4.77012590041537,1.22025583429631,6.41070111604149,0.0
+-6.92314832885029,2.86547174912175,8.7318591921429,0.0
+-7.56419504045061,4.94028695088613,8.94193466569898,0.0
+-8.63093667936286,4.27420502079032,9.27002100933965,0.0
+-8.99111149517962,5.10389362675139,11.7669513163404,0.0
+-9.99057638012989,7.87484596043453,12.4794035020575,0.0
+-10.3818782968703,8.84300238382604,13.7498993372395,0.0
+-11.0476824276485,9.44613324059569,13.502502696779,0.0
+-12.4344241411336,9.7051587072806,15.122117356301,0.0
+-13.6272942288038,10.419034297173,16.3289942692609,0.0
+-15.6202224973741,11.3788332186947,17.7367653402519,0.0
+-16.2922390869162,13.1516565107205,18.6939344625842,0.0
+-16.7159137753851,14.9076297530125,18.0246863680123,0.0
+-17.9501251836104,15.8533651031786,20.6826094110231,0.0
+-18.9898843433331,15.4331557188719,20.9101142300728,0.0
+-19.9085085832061,16.8542366574895,22.0721145894251,0.0
+-21.1466526906367,18.6785324946036,23.4977598550022,0.0
+-21.3675743976537,18.3208056358668,23.9121114020162,0.0
+-22.131396135351,20.7616214854992,24.1683442928502,0.0
+-23.1636319405886,21.1293492055766,25.2695476916398,0.0
+-24.1360762823224,21.7035705688561,27.9161820459799,0.0
+-25.3860721945057,23.3588003206916,27.8755285811136,0.0
+-27.2546274134222,24.9201403553661,28.9810564843169,0.0
+-28.84506179655,25.1681854417924,29.6749936376452,0.0
+-29.5648703289208,26.3989365511582,30.3290447782847,0.0
+-30.9473334996246,26.9014594846656,32.5460486188393,0.0
+-30.6333205969831,27.9838524246205,32.63981307435,0.0
+-31.1288464401156,28.0475385515748,34.4913247020856,0.0
+-32.9192755199859,30.8664195945179,35.0830004548623,0.0
+-33.5126863006676,31.7135111883763,35.963262553823,0.0
+-35.0626135760564,31.0828516323826,36.6124285861271,0.0
+-36.9524020008844,33.9722904006922,37.1501952879747,0.0
+-36.8859458347289,34.1203333509726,38.5388292299586,0.0
+-38.3029057124175,34.3054755372445,40.9141143587977,0.0
+-38.4946443648172,35.0506904211858,40.2609817417396,0.0
+-39.3511324722213,36.8560827180564,41.727133714198,0.0
+-41.7530250817957,37.750760484498,43.8282650147001,0.0
+-42.3509555518331,39.6673203525628,44.870018186039,0.0
+-42.244528415626,39.0788755186789,45.7949681462311,0.0
+-43.4340872344621,41.1905511155583,45.845002676574,0.0
+-45.6641407465385,42.0427797845611,46.7401996025804,0.0
+-46.3384022695138,42.4548123590202,47.6088880049696,0.0
+-47.2110032014801,43.9466417934236,49.2425272385793,0.0
+-48.5809364095671,45.2438778157098,49.7183364732398,0.0
+-48.8367941243197,45.8093002387181,51.8176699160374,0.0
+-49.3632780402121,47.2173660517208,51.910626136845,0.0
+-51.3304857529218,48.2292571877905,53.890768221757,0.0
+-52.6713212093626,48.0055890605259,54.2008024542977,0.0
+-53.0492822384067,50.9227140239038,55.8288448459473,0.0
+-53.5475407502719,50.399466736719,56.6653737429534,0.0
+-54.9713103390392,52.2174375936081,57.5892087088816,0.0
+-55.3824555656592,52.1624080567189,58.9048271122945,0.0
+-57.2392061954268,53.0027225338458,59.7821645733191,0.0
+-58.9104229152484,54.9892539015416,60.4219388424604,0.0
+-58.1824120199409,56.3542717685635,61.2574760633137,0.0
+-60.341564759061,56.7219808335487,61.0889655818828,0.0
+-61.7134385451754,58.6296683714548,63.962361122351,0.0
+-61.4485849309236,59.2722211731019,64.3931122256424,0.0
+-63.0320403297596,60.7473007790999,64.4546114078189,0.0
+-64.1041544802386,61.8740389113319,65.9649824945026,0.0
+-65.4439008347463,62.1787709650214,67.730652575126,0.0
+-65.3433374939793,63.9649079050439,67.2579323919331,0.0
+-66.1146018531461,64.8632381781096,69.639884237254,0.0
+-68.447747797377,65.4474036538936,69.6078430641903,0.0
+-69.294418370038,65.4181483428477,70.0720760179831,0.0
+-70.2849681665589,66.6959106050402,71.7102828712744,0.0
+-70.4280941800939,67.6086197232511,72.3531894807932,1.0
+-71.4902330755333,68.6959578032103,73.2224551274211,0.0
+-72.0927102365517,69.6523144077888,74.0072309103283,0.0
+-73.8239369250543,70.9727748471512,75.0623744862556,0.0
+-74.0048051089996,72.4351875425075,76.9734643262443,0.0
+-75.3894569624969,73.2005662031078,78.2395878278339,0.0
+-77.2087700153069,73.873044622658,79.2174499824634,1.0
+-78.3312007339737,74.3177101419062,80.4500892602852,0.0
+-78.0152851918014,76.0879642141581,81.7807815886821,1.0
+-79.6348564839424,76.2685721204187,82.516078707618,0.0
+-81.9637989522169,77.0024986767795,82.4661448054153,0.0
+-81.7384252154184,79.6668712051742,84.6703671298337,0.0
+-83.6602393343142,80.4913108235325,85.1069139287146,0.0
+-84.5817566138972,80.0267527066757,85.7126502390058,0.0
+-84.1500643836875,82.9742935504215,87.7889543910899,0.0
+-85.4066570769733,82.3347246270609,87.2294546291805,0.0
+-87.7637024688064,83.3797423673815,88.7428383996369,1.0
+-87.8532422216585,84.5508112461818,90.4419761299396,1.0
+-89.4596482634063,86.7455889667711,90.0173299461518,0.0
+-90.0512708441363,87.7184551425266,92.6086387957092,0.0
+-91.7774688537336,88.3362578318951,92.5323490472281,0.0
+-91.8941758249984,89.1756154247095,93.1997545484098,0.0
+-93.6063748804845,89.6466643235472,95.9725910658342,0.0
+-94.0398325903283,90.3812804437975,96.7702387180417,0.0
+-94.8766180342569,91.4799550730711,97.2352819068861,1.0
+-96.6522324427766,93.3217737988283,97.4817099617913,1.0
+-96.2874457012073,93.6988909891918,98.8166711820106,0.0
+-97.7197245583556,95.1054536364693,100.672788663657,0.0
+-98.1481566037241,95.1218250594043,100.732200143903,0.0
+-99.2071669041095,96.2150468128895,101.714216717689,0.0
+-100.936418858708,98.4045788153991,102.613459593007,1.0
+-102.443560747664,99.9196460296025,103.024089431854,1.0
+-103.253024963442,100.800612294368,104.827681828628,1.0
+-103.749612770423,101.647413918771,105.735965604858,1.0
+-104.328852482254,101.52632439617,106.261555245069,1.0
+-106.909854948268,103.417639882281,107.655384160196,0.0
+-107.146277874964,103.735504363015,109.010106082004,1.0
+-108.299680591253,105.221040236174,109.433289426765,1.0
+-108.483982549154,106.450953357466,110.204651921964,1.0
+-109.852652647883,107.528286824546,111.038226237901,1.0
+-110.908323481598,107.109110976433,113.280776112873,1.0
+-112.467924011291,108.685764539417,113.830693757837,1.0
+-113.205608139008,110.215691301935,114.433169300046,0.0
+-114.088065910725,111.855183307649,115.490962636794,1.0
+-114.932088811468,112.821293025765,116.840371472296,1.0
+-116.346270551874,112.158706089612,117.351225094843,0.0
+-117.586523469426,113.660859727683,119.786169957249,0.0
+-117.399020777901,115.140026386953,119.538050647704,0.0
+-119.742251706502,116.297893641214,121.055886377645,1.0
+-120.811568959087,116.442419636884,122.285416410297,1.0
+-121.87197183902,117.772559468586,123.134175983163,0.0
+-121.203532711179,118.217415486896,124.033131679149,1.0
+-122.799770233652,119.066944773458,125.957266439052,0.0
+-123.25939666012,121.092648631783,126.957392689029,1.0
+-124.665594116895,122.943504248815,126.726922729708,1.0
+-126.45629792163,123.680027227716,128.578446920242,1.0
+-126.478395127934,124.860787507164,129.382354701582,1.0
+-128.287481953834,124.904366629726,130.857704333381,1.0
+-129.634518688906,126.002157605703,130.664679201595,1.0
+-130.832869120258,126.312676682549,131.028091005298,1.0
+-131.824704198069,128.491040094009,132.438881937649,1.0
+-131.027748413386,128.666245091681,134.386819143277,1.0
+-133.251129013031,129.041770903105,135.442572078667,0.0
+-134.69199358524,131.797330031224,135.990306488139,1.0
+-134.269836322383,131.114035752663,137.455991445092,1.0
+-136.322562482609,133.458573512031,137.051219239798,1.0
+-137.181797709245,134.690639765394,139.80399233003,1.0
+-138.996622047529,135.735702581706,140.012525838337,1.0
+-139.215574081352,135.215412440418,141.384571880875,1.0
+-140.25643517594,136.693217029473,141.619094359002,1.0
+-141.091696435808,138.544186252854,143.447467807219,1.0
+-141.615279226692,139.247015471284,143.781551942357,1.0
+-143.738916186876,140.442376021513,145.690446407267,1.0
+-143.406661315583,141.611755086021,146.834954933569,1.0
+-145.539994351071,141.864759154179,146.821868669993,1.0
+-146.759792237469,143.72389809576,147.648189754648,1.0
+-147.742745321685,143.773042536024,148.335412957969,1.0
+-147.827110918347,144.891157822512,150.991195838727,1.0
+-149.385928352478,146.506476517532,151.284254745032,1.0
+-150.977776481931,146.021859562023,152.768419217875,1.0
+-150.312752816396,148.730155452661,152.038272157343,1.0
+-151.558966920835,148.800062079608,153.169105510546,1.0
+-152.600091618579,150.820711027947,154.708689228766,1.0
+-153.408136649531,151.844749166301,156.485789011581,1.0
+-155.851442988913,152.727616801779,157.051864933563,1.0
+-155.804122660613,153.080067542298,157.928538460402,1.0
+-157.073427576551,153.343498427848,158.067904946476,1.0
+-157.107508909122,154.26235452322,160.34811507316,1.0
+-159.484841744779,155.071935569765,161.331016311668,1.0
+-160.494393486944,156.832393661499,162.766796789861,1.0
+-160.993630097021,157.641448785465,162.16966187911,1.0
+-162.552759845799,159.539687734098,163.085996347117,1.0
+-163.564091719115,159.348960279159,164.704726412269,1.0
+-163.726283737347,161.928264713823,165.777638968284,1.0
+-165.875733565084,162.238166456333,167.748146433357,1.0
+-166.396130106934,163.168469688741,167.777541018587,1.0
+-166.362936492002,164.060990842406,168.793540054343,1.0
+-167.190830641697,164.69633527615,169.716276093277,1.0
+-169.856337102559,165.948746195011,171.767907484416,1.0
+-170.437574285639,166.585793448766,172.423441930238,1.0
+-171.364589341033,167.713387609531,172.424258150602,1.0
+-171.719858914268,168.568258270241,174.186728478313,1.0
+-172.222432925774,169.341495037779,175.014592418557,1.0
+-174.925024864504,170.009200154178,175.823166747739,1.0
+-174.440609108192,172.585679891303,177.497935948773,1.0
+-175.685134632762,172.597439708769,178.865842811381,1.0
+-177.037735766697,174.337033018417,178.526154196112,1.0
+-177.695754636999,175.158079880788,179.090845324338,1.0
+-178.530965125958,175.998360662267,180.756244303945,1.0
+-179.33859605694,177.055963520977,182.537104759468,1.0
+-181.051026538749,178.749994896864,182.511752447157,1.0
+-181.910324655635,178.191126997934,183.205490333945,1.0
+-182.658814367512,179.821227820741,184.077256331919,1.0
+-184.92643232869,181.744602320009,185.895789796461,1.0
+-185.336918062898,181.363095955327,186.129937394056,1.0
+-185.267288986526,182.500514904395,188.252778726315,1.0
+-186.872953708712,184.965877299035,188.828219832583,1.0
+-188.918306879815,185.051974487609,190.899813628421,1.0
+-189.763521615511,185.68824128269,190.220429056165,1.0
+-190.644174012638,187.750343044182,191.493983126918,1.0
+-190.321219894606,187.817497818665,193.40572238254,1.0
+-192.79354360347,188.86697228881,193.159805548127,1.0
+-192.025564246072,189.147494777077,194.034017243282,1.0
+-193.32579637306,190.629543307974,195.316361684183,1.0
+-194.09803396593,192.184663556046,196.170525930209,1.0
+-196.951286736902,192.950569139836,198.865411666979,1.0
+-197.273285018295,193.455170048879,199.984457526702,1.0
+-197.367054178611,194.765784862729,199.831724804912,1.0

data/spec/data/normal.csv

@@ -0,0 +1,30 @@
+ROLL,UNEM,HGRAD,INC
+5501,8.1,9552,1923
+5945,7,9680,1961
+6629,7.3,9731,1979
+7556,7.5,11666,2030
+8716,7,14675,2112
+9369,6.4,15265,2192
+9920,6.5,15484,2235
+10167,6.4,15723,2351
+11084,6.3,16501,2411
+12504,7.7,16890,2475
+13746,8.2,17203,2524
+13656,7.5,17707,2674
+13850,7.4,18108,2833
+14145,8.2,18266,2863
+14888,10.1,19308,2839
+14991,9.2,18224,2898
+14836,7.7,18997,3123
+14478,5.7,19505,3195
+14539,6.5,19800,3239
+14395,7.5,19546,3129
+14599,7.3,19117,3100
+14969,9.2,18774,3008
+15107,10.1,17813,2983
+14831,7.5,17304,3069
+15081,8.8,16756,3151
+15127,9.1,16749,3127
+15856,8.8,16925,3179
+15938,7.8,17231,3207
+16081,7,16816,3345

data/spec/logistic_spec.rb

@@ -0,0 +1,37 @@
+require 'spec_helper.rb'
+
+describe Statsample::GLM::Logistic do
+  context "IRLS algorithm" do
+    before do
+      @data_set = Statsample::CSV.read "spec/data/logistic.csv"
+      @glm  = Statsample::GLM.compute @data_set, 'y', :logistic, {constant: 1}
+    end
+
+    it "reports correct coefficients as an array" do
+      expect_similar_vector(@glm.coefficients,[-0.312493754568903,
+        2.28671333346264,0.675603176233325])
+
+    end
+
+    it "reports correct coefficients as a hash" do
+      expect_similar_hash(@glm.coefficients(:hash), {:constant => 0.675603176233325,
+        :x1 => -0.312493754568903, :x2 => 2.28671333346264})
+    end
+  end
+
+  context "MLE algorithm" do
+    before do
+      @data_set = Statsample::CSV.read("spec/data/logistic_mle.csv")
+      @glm      = Statsample::GLM.compute @data_set,'y', :logistic, {constant: 1, algorithm: :mle}
+    end
+
+    it "reports correct regression values as an array" do
+      expect(@glm.log_likelihood).to be_within(0.001).of(-38.8669)
+
+      expect_similar_vector(@glm.coefficients, [0.3270, 0.8147, -0.4031,-5.3658],0.001)
+      expect_similar_vector(@glm.standard_error, [0.4390, 0.4270, 0.3819,1.9045],0.001)
+
+      expect(@glm.iterations).to eq(7)
+    end
+  end
+end