search

salesforce / TransmogrifAI

TransmogrifAI (pronounced trăns-mŏgˈrə-fī) is an AutoML library for building modular, reusable, strongly typed machine learning workflows on Apache Spark with minimal hand-tuning
https://transmogrif.ai
BSD 3-Clause "New" or "Revised" License
2.24k stars 393 forks source link

Not able to reproduce results #371

Closed pvn25 closed 5 years ago

pvn25 commented 5 years ago

Describe the bug Unable to reproduce results when running multiple times

To Reproduce

%classpath add mvn com.salesforce.transmogrifai transmogrifai-core_2.11 0.5.3
%classpath add mvn org.apache.spark spark-mllib_2.11 2.3.2

import org.apache.spark.SparkConf
import org.apache.spark.sql.SparkSession
import org.apache.spark.SparkContext
import org.apache.spark.sql.functions.udf
import com.salesforce.op._
import com.salesforce.op.features._
import com.salesforce.op.features.types._
import com.salesforce.op.stages.impl.classification._
import com.salesforce.op.evaluators.Evaluators

val conf = new SparkConf().setMaster("local[*]").setAppName("TitanicPrediction")
implicit val spark = SparkSession.builder.config(conf).getOrCreate()

import spark.implicits._ // Needed for Encoders for the Passenger case class
import com.salesforce.op.readers.DataReaders

case class Passenger(
  id: Int,
  survived: Double,
  pClass: Option[Int],
  name: Option[String],
  sex: Option[String],
  age: Option[Double],
  sibSp: Option[Int],
  parCh: Option[Int],
  ticket: Option[String],
  fare: Option[Double],
  cabin: Option[String],
  embarked: Option[String]
)

val pathToData = "titanic_train.csv"
val passengersData = DataReaders.Simple.csvCase[Passenger](path = Option(pathToData)).readDataset().toDF()

val (survived, predictors) = FeatureBuilder.fromDataFrame[RealNN](passengersData, response = "survived")
val features = predictors.transmogrify()

import com.salesforce.op.stages.impl.classification.{BinaryClassificationModelsToTry => BMT}
import com.salesforce.op.stages.impl.tuning.DataSplitter
import org.apache.spark.ml.tuning.ParamGridBuilder
import com.salesforce.op.stages.impl.selector.{DefaultSelectorParams, ModelSelectorFactory, ModelSelector}

val randomSeed = 300L
val splitter = DataSplitter(seed = randomSeed)

val rf = new OpRandomForestClassifier().setSeed(randomSeed) // set the seed for RF classifier
val rfParams = new ParamGridBuilder()
    .addGrid(rf.maxDepth, DefaultSelectorParams.MaxDepth)
    .addGrid(rf.impurity, DefaultSelectorParams.ImpurityClass)
    .addGrid(rf.maxBins, DefaultSelectorParams.MaxBin)
    .addGrid(rf.minInfoGain, DefaultSelectorParams.MinInfoGain)
    .addGrid(rf.minInstancesPerNode, DefaultSelectorParams.MinInstancesPerNode)
    .addGrid(rf.numTrees, DefaultSelectorParams.MaxTrees)
    .addGrid(rf.subsamplingRate, DefaultSelectorParams.SubsampleRate)
    .build()

val modelsAndParameters = Seq(rf -> rfParams)

val prediction = BinaryClassificationModelSelector
    .withCrossValidation(splitter = Option(splitter), seed = randomSeed, modelsAndParameters = modelsAndParameters)
    .setInput(survived, features)
    .getOutput()

val model = new OpWorkflow().setInputDataset(passengersData).setResultFeatures(prediction).train()
val evaluator = Evaluators.BinaryClassification().setLabelCol(survived).setPredictionCol(prediction)

 // Manifest the result features of the workflow
println("Scoring the model")
val (dataframe, metrics) = model.scoreAndEvaluate(evaluator = evaluator)

Expected behavior I'm expecting same results ever run

hzrt123 commented 5 years ago

Can you post some 'different' results here ?

pvn25 commented 5 years ago

for example, one run gave me this: 


Scoring the model

{
  "Precision" : 0.9777777777777777,
  "Recall" : 0.1286549707602339,
  "F1" : 0.22739018087855295,
  "AuROC" : 0.8593615185504746,
  "AuPR" : 0.8143447031549991,
  "Error" : 0.33557800224466894,
  "TP" : 44.0,
  "TN" : 548.0,
  "FP" : 1.0,
  "FN" : 298.0,
  "thresholds" : [ 0.5466022294774278, 0.5303642833069607, 0.5245191408903661, 0.5201384571600074, 0.5147680307359522, 0.5131403928517401, 0.5099803968936437, 0.5082282399764858, 0.5021277010233705, 0.498784546668842, 0.49613988012903126, 0.4927756201229488, 0.4913754818953017, 0.4890923951961965, 0.48700268545780523, 0.4847073088003347, 0.48301488798376513, 0.48140515619018415, 0.47927655245500744, 0.4772064790834232, 0.47619984170872826, 0.4734296459112426, 0.4722520047259987, 0.46975015297424816, 0.4686017322329081, 0.4674804441714503, 0.4662855524698201, 0.46492523585844386, 0.4643706813770827, 0.4635877761231427, 0.4631089759526395, 0.4620111534823743, 0.4593255099407294, 0.45735297384531653, 0.45662222016619836, 0.45519713563186104, 0.45461937900220656, 0.4529837176958532, 0.4517634596067406, 0.4507631951621246, 0.4489878177998871, 0.4468376179994136, 0.44439917906639687, 0.4430875418076844, 0.44208478811535484, 0.44051258027134865, 0.4394656225848582, 0.43565513110577975, 0.43425129641326615, 0.4309741932991811, 0.4280531880144576, 0.42439366443514487, 0.42160925471946054, 0.4148731309444286, 0.40945720576585615, 0.4051702165569533, 0.3969027409863694, 0.39343398622395775, 0.3914599100860711, 0.3895084311888095, 0.38617875820015185, 0.38290356386988456, 0.38023945391691355, 0.37749642667688177, 0.3763706434192269, 0.37220948245772595, 0.3710271032932065, 0.37005475463204446, 0.3690369206086107, 0.36583083797159377, 0.3633016281457233, 0.36187144980735125, 0.3589850095122891, 0.3567832681204717, 0.35437738026439425, 0.3524158605761189, 0.3502393583339664, 0.3479873334062081, 0.3476087679962962, 0.3466844252404932, 0.3456340781154119, 0.3443007111467158, 0.34211534768393553, 0.34105177253969865, 0.34009894055616874, 0.3391992467003549, 0.3384202638524627, 0.3378066012273228, 0.337041881471723, 0.33623261731729476, 0.3349810240934841, 0.3337103623235011, 0.33281146681936497, 0.33173744799570837, 0.33026236144417503, 0.3296008496214213, 0.3288670281902727, 0.3274465109138944, 0.3266817114026985, 0.3262911996372801, 0.3253587699946586, 0.32476950877740696, 0.32391572625922727, 0.32287177772035525, 0.3217121534533052, 0.3206263783463074, 0.3198419470735569, 0.31815820832875125, 0.3170629221493905, 0.31523407587434926, 0.31348939585049185, 0.31177454002063887, 0.31100299438093437, 0.31020305711996177, 0.30802655487780933, 0.3058093320836948, 0.30330335075010495, 0.3004107257588848, 0.2978779099009353, 0.29234331870596675 ],
  "precisionByThreshold" : [ 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.972972972972973, 0.9761904761904762, 0.9787234042553191, 0.9811320754716981, 0.9827586206896551, 0.9841269841269841, 0.9705882352941176, 0.958904109589041, 0.9358974358974359, 0.927710843373494, 0.9318181818181818, 0.925531914893617, 0.9191919191919192, 0.9038461538461539, 0.908256880733945, 0.8947368421052632, 0.8907563025210085, 0.8951612903225806, 0.8914728682170543, 0.8955223880597015, 0.8992805755395683, 0.8958333333333334, 0.8791946308724832, 0.8846153846153846, 0.8827160493827161, 0.8802395209580839, 0.8837209302325582, 0.8757062146892656, 0.8743169398907104, 0.8723404255319149, 0.8608247422680413, 0.8592964824120602, 0.8454106280193237, 0.8450704225352113, 0.8493150684931506, 0.8482142857142857, 0.8434782608695652, 0.8425531914893617, 0.8423236514522822, 0.8373983739837398, 0.8286852589641435, 0.8262548262548263, 0.8102189781021898, 0.8035714285714286, 0.7894736842105263, 0.7808219178082192, 0.7573770491803279, 0.744408945686901, 0.7452830188679245, 0.7368421052631579, 0.7378048780487805, 0.7357357357357357, 0.7337278106508875, 0.7259475218658892, 0.7277936962750716, 0.7175141242937854, 0.7186629526462396, 0.7142857142857143, 0.7127371273712737, 0.704, 0.7078947368421052, 0.7072538860103627, 0.7007672634271099, 0.696969696969697, 0.6907730673316709, 0.675609756097561, 0.672289156626506, 0.6643026004728132, 0.6596736596736597, 0.6507936507936508, 0.6481069042316259, 0.6461538461538462, 0.6363636363636364, 0.6316916488222698, 0.625, 0.6171548117154811, 0.6107660455486542, 0.6040816326530613, 0.6032064128256514, 0.5968379446640316, 0.5909980430528375, 0.584942084942085, 0.5798479087452472, 0.5749063670411985, 0.5703703703703704, 0.5669724770642202, 0.552212389380531, 0.540587219343696, 0.535958904109589, 0.5314091680814941, 0.5251677852348994, 0.5190713101160862, 0.5163934426229508, 0.5113636363636364, 0.507223113964687, 0.5015873015873016, 0.49453978159126366, 0.4885145482388974, 0.4878419452887538, 0.48493975903614456, 0.48059701492537316, 0.47584187408491946, 0.4704184704184704, 0.46403385049365303, 0.44489247311827956, 0.4225, 0.4200743494423792, 0.4166666666666667, 0.407673860911271, 0.4052443384982122, 0.4007050528789659, 0.3974358974358974, 0.3855693348365276, 0.3838383838383838 ],
  "recallByThreshold" : [ 0.014619883040935672, 0.029239766081871343, 0.04678362573099415, 0.06140350877192982, 0.07602339181286549, 0.0935672514619883, 0.10526315789473684, 0.11988304093567251, 0.13450292397660818, 0.15204678362573099, 0.16666666666666666, 0.18128654970760233, 0.19298245614035087, 0.2046783625730994, 0.2134502923976608, 0.22514619883040934, 0.23976608187134502, 0.2543859649122807, 0.26608187134502925, 0.27485380116959063, 0.2894736842105263, 0.2982456140350877, 0.30994152046783624, 0.32456140350877194, 0.3362573099415205, 0.3508771929824561, 0.3654970760233918, 0.37719298245614036, 0.3830409356725146, 0.40350877192982454, 0.41812865497076024, 0.4298245614035088, 0.4444444444444444, 0.45321637426900585, 0.4678362573099415, 0.47953216374269003, 0.48830409356725146, 0.5, 0.5116959064327485, 0.5263157894736842, 0.543859649122807, 0.5555555555555556, 0.5672514619883041, 0.5789473684210527, 0.5935672514619883, 0.6023391812865497, 0.6081871345029239, 0.6257309941520468, 0.6491228070175439, 0.6578947368421053, 0.6578947368421053, 0.6666666666666666, 0.6754385964912281, 0.6812865497076024, 0.6929824561403509, 0.695906432748538, 0.7076023391812866, 0.716374269005848, 0.7251461988304093, 0.7280701754385965, 0.7426900584795322, 0.7426900584795322, 0.7543859649122807, 0.7602339181286549, 0.7690058479532164, 0.7719298245614035, 0.7865497076023392, 0.7982456140350878, 0.8011695906432749, 0.8070175438596491, 0.8099415204678363, 0.8099415204678363, 0.8157894736842105, 0.8216374269005848, 0.827485380116959, 0.8391812865497076, 0.8508771929824561, 0.8596491228070176, 0.8596491228070176, 0.8625730994152047, 0.8625730994152047, 0.8625730994152047, 0.8625730994152047, 0.8654970760233918, 0.8801169590643275, 0.8830409356725146, 0.8830409356725146, 0.8859649122807017, 0.8918128654970761, 0.8976608187134503, 0.9005847953216374, 0.9035087719298246, 0.9122807017543859, 0.9152046783625731, 0.9152046783625731, 0.9152046783625731, 0.9152046783625731, 0.9152046783625731, 0.9210526315789473, 0.9210526315789473, 0.9239766081871345, 0.9239766081871345, 0.9269005847953217, 0.9327485380116959, 0.9385964912280702, 0.9415204678362573, 0.9415204678362573, 0.9502923976608187, 0.9532163742690059, 0.9619883040935673, 0.9678362573099415, 0.9883040935672515, 0.9912280701754386, 0.9941520467836257, 0.9941520467836257, 0.9941520467836257, 0.9970760233918129, 0.9970760233918129, 1.0, 1.0 ],
  "falsePositiveRateByThreshold" : [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0036429872495446266, 0.00546448087431694, 0.009107468123861567, 0.01092896174863388, 0.01092896174863388, 0.012750455373406194, 0.014571948998178506, 0.018214936247723135, 0.018214936247723135, 0.02185792349726776, 0.023679417122040074, 0.023679417122040074, 0.025500910746812388, 0.025500910746812388, 0.025500910746812388, 0.0273224043715847, 0.03278688524590164, 0.03278688524590164, 0.03460837887067395, 0.03642987249544627, 0.03642987249544627, 0.04007285974499089, 0.04189435336976321, 0.04371584699453552, 0.04918032786885246, 0.051001821493624776, 0.058287795992714025, 0.060109289617486336, 0.060109289617486336, 0.061930783242258654, 0.06557377049180328, 0.06739526411657559, 0.0692167577413479, 0.07285974499089254, 0.07832422586520947, 0.08196721311475409, 0.0947176684881603, 0.10018214936247723, 0.1092896174863388, 0.11657559198542805, 0.13479052823315119, 0.14571948998178508, 0.14754098360655737, 0.15482695810564662, 0.15664845173041894, 0.16029143897996356, 0.16393442622950818, 0.17122040072859745, 0.17304189435336975, 0.18214936247723132, 0.18397085610200364, 0.1894353369763206, 0.1930783242258652, 0.20218579234972678, 0.20218579234972678, 0.2058287795992714, 0.21311475409836064, 0.2185792349726776, 0.22586520947176686, 0.24225865209471767, 0.24772313296903462, 0.2586520947176685, 0.2659380692167577, 0.28051001821493626, 0.2877959927140255, 0.29326047358834245, 0.30601092896174864, 0.3132969034608379, 0.3224043715846995, 0.3333333333333333, 0.3424408014571949, 0.3533697632058288, 0.36065573770491804, 0.37158469945355194, 0.3806921675774135, 0.39162112932604737, 0.40255009107468126, 0.4134790528233151, 0.4225865209471767, 0.42987249544626593, 0.46083788706739526, 0.48451730418943534, 0.4936247723132969, 0.5027322404371585, 0.5154826958105647, 0.5282331511839709, 0.5373406193078324, 0.5482695810564663, 0.5591985428051002, 0.5719489981785064, 0.5901639344262295, 0.6083788706739527, 0.6138433515482696, 0.6229508196721312, 0.6338797814207651, 0.6520947176684881, 0.668488160291439, 0.692167577413479, 0.7522768670309654, 0.8415300546448088, 0.8524590163934426, 0.8670309653916212, 0.8998178506375227, 0.9089253187613844, 0.9289617486338798, 0.941712204007286, 0.9927140255009107, 1.0 ]
}

and another run this:

Scoring the model

{
  "Precision" : 0.9629629629629629,
  "Recall" : 0.38011695906432746,
  "F1" : 0.5450733752620545,
  "AuROC" : 0.8656328891445371,
  "AuPR" : 0.821126645714924,
  "Error" : 0.24354657687991021,
  "TP" : 130.0,
  "TN" : 544.0,
  "FP" : 5.0,
  "FN" : 212.0,
  "thresholds" : [ 0.5752840846942466, 0.5526070646032224, 0.5420765070075221, 0.5367302044322335, 0.5345010436604243, 0.5320079184251643, 0.5295915579684766, 0.5262489662235466, 0.5233677321138692, 0.521428366144255, 0.5194740585205723, 0.5168775486825998, 0.5153589598871189, 0.5127486259339882, 0.5113351246913801, 0.5089019890607162, 0.5074225902703408, 0.5059154804562118, 0.5039843406690001, 0.5028946361207455, 0.5020029784443123, 0.49825913710295006, 0.49532596120946787, 0.4937777436952681, 0.4928806549180198, 0.48963353206424054, 0.4879682674331242, 0.4859463193647364, 0.48475953219161577, 0.48226451181139746, 0.48003171531832306, 0.4778878960976999, 0.47584210434216734, 0.4745724300961103, 0.47296673934787214, 0.47134216988748173, 0.46927546500722, 0.468056148836743, 0.4657346920488667, 0.4629518939986527, 0.4591137493162928, 0.45730684613366285, 0.4523846411213221, 0.44437729581836544, 0.40576967002473585, 0.3994099422405744, 0.39398760334600336, 0.391089947038605, 0.3889590423762536, 0.3852289826563482, 0.38171364163071014, 0.37781899772284516, 0.37632555592913913, 0.37274300375499064, 0.3705700846971972, 0.36767604670376813, 0.3652737694533141, 0.36220764450705123, 0.358920713127506, 0.35623758845728704, 0.354495040819924, 0.3533957522129421, 0.3510293528743551, 0.3489257181930615, 0.34776957732421204, 0.3450318161091101, 0.34416173038323616, 0.34303470215657594, 0.3414833449134912, 0.3402012664743046, 0.33875952086408373, 0.337164800432738, 0.33582839483905486, 0.335271958635734, 0.33224960824237804, 0.33105205100366136, 0.33052020714533115, 0.32914438208817137, 0.3278424238571525, 0.3265789890065414, 0.32580345810354244, 0.32483845794093164, 0.32317027121133746, 0.3226636153874957, 0.3216023252016912, 0.3201888337071269, 0.31904372739398923, 0.3178209140812825, 0.3169450761061862, 0.3155225399199258, 0.3148151810755495, 0.31437188859659015, 0.3132700966196273, 0.311857565265024, 0.3104774652960405, 0.30917408431124355, 0.3064589066168682, 0.30380349653587835, 0.3012306395811664, 0.2949576234620644, 0.2884392378427396 ],
  "precisionByThreshold" : [ 0.8333333333333334, 0.9333333333333333, 0.9523809523809523, 0.9629629629629629, 0.9714285714285714, 0.9761904761904762, 0.9791666666666666, 0.9629629629629629, 0.967741935483871, 0.9710144927536232, 0.9733333333333334, 0.9753086419753086, 0.9659090909090909, 0.9578947368421052, 0.9603960396039604, 0.9626168224299065, 0.9557522123893806, 0.957983193277311, 0.96, 0.9618320610687023, 0.9635036496350365, 0.9444444444444444, 0.9205298013245033, 0.9177215189873418, 0.9096385542168675, 0.8920454545454546, 0.8852459016393442, 0.8795811518324608, 0.8592964824120602, 0.8585365853658536, 0.8584905660377359, 0.8486238532110092, 0.8454935622317596, 0.8381742738589212, 0.8299595141700404, 0.8171206225680934, 0.8068181818181818, 0.7896678966789668, 0.7841726618705036, 0.7627118644067796, 0.7574750830564784, 0.752442996742671, 0.7412140575079872, 0.746875, 0.7515337423312883, 0.7469879518072289, 0.7411764705882353, 0.7398843930635838, 0.7329545454545454, 0.729050279329609, 0.7252747252747253, 0.7189189189189189, 0.7180851063829787, 0.7094240837696335, 0.6936708860759494, 0.6890547263681592, 0.6838235294117647, 0.678743961352657, 0.6738095238095239, 0.6713615023474179, 0.6651270207852193, 0.6554809843400448, 0.6512141280353201, 0.6456521739130435, 0.6437768240343348, 0.6350210970464135, 0.6286307053941909, 0.61, 0.6023622047244095, 0.5949612403100775, 0.5874524714828897, 0.5786516853932584, 0.5730129390018485, 0.5664845173041895, 0.5512367491166078, 0.5454545454545454, 0.5358361774744027, 0.5277310924369748, 0.5257903494176372, 0.5179738562091504, 0.5121163166397416, 0.5088, 0.49767441860465117, 0.4938650306748466, 0.48714069591527986, 0.48353293413173654, 0.4807121661721068, 0.4385733157199472, 0.43246753246753245, 0.42528735632183906, 0.4247787610619469, 0.4228356336260979, 0.41370869033047736, 0.41019417475728154, 0.40816326530612246, 0.39605110336817656, 0.39285714285714285, 0.38971428571428574, 0.3873159682899207, 0.3847019122609674, 0.3838383838383838 ],
  "recallByThreshold" : [ 0.014619883040935672, 0.04093567251461988, 0.05847953216374269, 0.07602339181286549, 0.09941520467836257, 0.11988304093567251, 0.13742690058479531, 0.15204678362573099, 0.17543859649122806, 0.195906432748538, 0.2134502923976608, 0.2309941520467836, 0.24853801169590642, 0.26608187134502925, 0.28362573099415206, 0.30116959064327486, 0.3157894736842105, 0.3333333333333333, 0.3508771929824561, 0.3684210526315789, 0.38596491228070173, 0.39766081871345027, 0.4064327485380117, 0.4239766081871345, 0.4415204678362573, 0.4590643274853801, 0.47368421052631576, 0.49122807017543857, 0.5, 0.5146198830409356, 0.5321637426900585, 0.5409356725146199, 0.5760233918128655, 0.5906432748538012, 0.5994152046783626, 0.6140350877192983, 0.6228070175438597, 0.6257309941520468, 0.6374269005847953, 0.6578947368421053, 0.6666666666666666, 0.6754385964912281, 0.6783625730994152, 0.6988304093567251, 0.716374269005848, 0.7251461988304093, 0.7368421052631579, 0.7485380116959064, 0.7543859649122807, 0.7631578947368421, 0.7719298245614035, 0.7777777777777778, 0.7894736842105263, 0.7923976608187134, 0.8011695906432749, 0.8099415204678363, 0.8157894736842105, 0.8216374269005848, 0.827485380116959, 0.8362573099415205, 0.8421052631578947, 0.8567251461988304, 0.8625730994152047, 0.868421052631579, 0.8771929824561403, 0.8801169590643275, 0.8859649122807017, 0.8918128654970761, 0.8947368421052632, 0.8976608187134503, 0.9035087719298246, 0.9035087719298246, 0.9064327485380117, 0.9093567251461988, 0.9122807017543859, 0.9122807017543859, 0.9181286549707602, 0.9181286549707602, 0.9239766081871345, 0.9269005847953217, 0.9269005847953217, 0.9298245614035088, 0.9385964912280702, 0.9415204678362573, 0.9415204678362573, 0.9444444444444444, 0.9473684210526315, 0.9707602339181286, 0.9736842105263158, 0.9736842105263158, 0.9824561403508771, 0.9853801169590644, 0.9883040935672515, 0.9883040935672515, 0.9941520467836257, 0.9970760233918129, 0.9970760233918129, 0.9970760233918129, 1.0, 1.0, 1.0 ],
  "falsePositiveRateByThreshold" : [ 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0018214936247723133, 0.0036429872495446266, 0.0036429872495446266, 0.0036429872495446266, 0.0036429872495446266, 0.0036429872495446266, 0.00546448087431694, 0.007285974499089253, 0.007285974499089253, 0.007285974499089253, 0.009107468123861567, 0.009107468123861567, 0.009107468123861567, 0.009107468123861567, 0.009107468123861567, 0.014571948998178506, 0.02185792349726776, 0.023679417122040074, 0.0273224043715847, 0.03460837887067395, 0.03825136612021858, 0.04189435336976321, 0.051001821493624776, 0.052823315118397086, 0.0546448087431694, 0.060109289617486336, 0.06557377049180328, 0.07103825136612021, 0.07650273224043716, 0.08561020036429873, 0.09289617486338798, 0.10382513661202186, 0.1092896174863388, 0.12750455373406194, 0.13296903460837886, 0.1384335154826958, 0.14754098360655737, 0.14754098360655737, 0.14754098360655737, 0.15300546448087432, 0.16029143897996356, 0.16393442622950818, 0.17122040072859745, 0.1766848816029144, 0.18214936247723132, 0.1894353369763206, 0.1930783242258652, 0.20218579234972678, 0.2204007285974499, 0.22768670309653916, 0.23497267759562843, 0.24225865209471767, 0.2495446265938069, 0.2550091074681239, 0.2641165755919854, 0.28051001821493626, 0.2877959927140255, 0.29690346083788705, 0.302367941712204, 0.3151183970856102, 0.32604735883424407, 0.3551912568306011, 0.3679417122040073, 0.3806921675774135, 0.39526411657559196, 0.4098360655737705, 0.4207650273224044, 0.4335154826958106, 0.46265938069216755, 0.47358834244080145, 0.49544626593806923, 0.51183970856102, 0.5191256830601093, 0.5373406193078324, 0.5500910746812386, 0.5591985428051002, 0.5901639344262295, 0.6010928961748634, 0.6174863387978142, 0.6284153005464481, 0.6375227686703097, 0.7741347905282332, 0.7959927140255009, 0.819672131147541, 0.8287795992714025, 0.8378870673952641, 0.8724954462659381, 0.8852459016393442, 0.8979963570127505, 0.9471766848816029, 0.9599271402550091, 0.9726775956284153, 0.9854280510018215, 0.9963570127504554, 1.0 ]
}
hzrt123 commented 5 years ago

In theory, the results could be a bit different due to some random steps in the algorithms.

Your demo results show obvious difference, so I guess your data set is too small so that the algorithms don't have a stable result.

Maybe you can use some bigger data set to test.