Polynomial Logistic Regression weird convergence issue? Bug?
Hi!
I'm wondering if anyone could share some insight into this problem. I'm trying to run an evolutionary search of parameters for the mentioned operator, but it won't converge.
I've been trying to find out why, but ran into some trouble pinning it down. Take the sonar data example proposed by the logistic regression operator:
2) Then I changed max iterations to 1 and started going up. 291 max iterations runs fine, again, almost instantly. 292 max iterations, one minute going already... won't converge.
I'm not a doctor in machine learning algorithms but this looks like a bug to me ;D...
Any ideas of how to cope with this while it gets fixed?
Thanks! I hope the bug report helps.
I'm wondering if anyone could share some insight into this problem. I'm trying to run an evolutionary search of parameters for the mentioned operator, but it won't converge.
I've been trying to find out why, but ran into some trouble pinning it down. Take the sonar data example proposed by the logistic regression operator:
<?xml version="1.0" encoding="UTF-8" standalone="no"?>1) I simply changed to polynomial and set degree at 12. This runs instantly on my PC, but when changing degree to 13, it won't converge not even past 5 minutes.
<process version="5.3.015">
<context>
<input/>
<output/>
<macros/>
</context>
<operator activated="true" class="process" compatibility="5.3.015" expanded="true" name="Process">
<process expanded="true">
<operator activated="true" class="retrieve" compatibility="5.3.015" expanded="true" height="60" name="Sonar" width="90" x="380" y="120">
<parameter key="repository_entry" value="//Samples/data/Sonar"/>
</operator>
<operator activated="true" class="split_validation" compatibility="5.3.015" expanded="true" height="112" name="Validation" width="90" x="514" y="120">
<process expanded="true">
<operator activated="true" class="logistic_regression" compatibility="5.3.015" expanded="true" height="94" name="Logistic Regression" width="90" x="112" y="30">
<parameter key="kernel_type" value="polynomial"/>
<parameter key="kernel_degree" value="12.0"/>
</operator>
<connect from_port="training" to_op="Logistic Regression" to_port="training set"/>
<connect from_op="Logistic Regression" from_port="model" to_port="model"/>
<portSpacing port="source_training" spacing="0"/>
<portSpacing port="sink_model" spacing="0"/>
<portSpacing port="sink_through 1" spacing="0"/>
</process>
<process expanded="true">
<operator activated="true" class="apply_model" compatibility="5.3.015" expanded="true" height="76" name="Apply Model" width="90" x="45" y="30">
<list key="application_parameters"/>
</operator>
<operator activated="true" class="performance" compatibility="5.3.015" expanded="true" height="76" name="Performance" width="90" x="179" y="30"/>
<connect from_port="model" to_op="Apply Model" to_port="model"/>
<connect from_port="test set" to_op="Apply Model" to_port="unlabelled data"/>
<connect from_op="Apply Model" from_port="labelled data" to_op="Performance" to_port="labelled data"/>
<connect from_op="Performance" from_port="performance" to_port="averagable 1"/>
<portSpacing port="source_model" spacing="0"/>
<portSpacing port="source_test set" spacing="0"/>
<portSpacing port="source_through 1" spacing="0"/>
<portSpacing port="sink_averagable 1" spacing="0"/>
<portSpacing port="sink_averagable 2" spacing="0"/>
</process>
</operator>
<connect from_op="Sonar" from_port="output" to_op="Validation" to_port="training"/>
<connect from_op="Validation" from_port="model" to_port="result 1"/>
<connect from_op="Validation" from_port="averagable 1" to_port="result 2"/>
<portSpacing port="source_input 1" spacing="0"/>
<portSpacing port="sink_result 1" spacing="90"/>
<portSpacing port="sink_result 2" spacing="18"/>
<portSpacing port="sink_result 3" spacing="0"/>
</process>
</operator>
</process>
2) Then I changed max iterations to 1 and started going up. 291 max iterations runs fine, again, almost instantly. 292 max iterations, one minute going already... won't converge.
I'm not a doctor in machine learning algorithms but this looks like a bug to me ;D...
Any ideas of how to cope with this while it gets fixed?
Thanks! I hope the bug report helps.
Find more posts tagged with
Sort by:
1 - 5 of
51
mafern76
New Altair Community Member
OPPS. with same scenario I tried anova and epach with 1000000 degree and it runs instantly.
Altair EmployeeHi,
this looks weird indeed. I have created an interal ticket for the issue. Thank you for reporting it!
Regards,
Marco
this looks weird indeed. I have created an interal ticket for the issue. Thank you for reporting it!
Regards,
Marco
New Altair Community Member
OPThanks Marco..
An update on anova and epach, on a real-case scenario (1000/1000 cases, 17 predictors, evo search) anova also does get stuck immediately. Epach runs but seems to be the slowest, by far, of all the log. reg. kernel types.
This anova runs slowly but at least it runs:
An update on anova and epach, on a real-case scenario (1000/1000 cases, 17 predictors, evo search) anova also does get stuck immediately. Epach runs but seems to be the slowest, by far, of all the log. reg. kernel types.
This anova runs slowly but at least it runs:
<parameter key="LR_ANOVA_MOD.kernel_gamma" value="[0.00001;1]"/>RM gets a little bit unresponsive though. Maybe something overloads something else, somewhere, somehow?.. some combinations of parameters seem to be growing the problem exponentially, sometimes to a point of no return.
<parameter key="LR_ANOVA_MOD.kernel_degree" value="[1;2]"/>
<parameter key="LR_ANOVA_MOD.C" value="[0;10]"/>
<parameter key="LR_ANOVA_MOD.convergence_epsilon" value="[0.1;0.5]"/>
New Altair Community Member
OPMaybe this short run log can help: as you can see on a plot, time between validations varies a lot.
Oh and by the way, I think this is the first time I've seen this... C takes values beyond the '10' limit, maybe it's related, or maybe just another bug, I don't know.
Oh and by the way, I think this is the first time I've seen this... C takes values beyond the '10' limit, maybe it's related, or maybe just another bug, I don't know.
time memo c gamma degree conv_epsilon perfo perfo_dev
38150.0 1.309247264E9 5.708774309130852 0.4815469360780778 1.6894985156297109 0.19529838321505755 0.6884993281209857 0.009104510275475342
40917.0 9.19223E8 0.34267737363115836 0.5404634943325216 1.1315150563914869 0.3353537248701405 0.7190726762798834 0.014539775408387932
41967.0 9.52729576E8 4.045221107690375 0.5418650318132358 1.15506380864328 0.34045601477565923 0.7129182553289294 0.007991655337273243
45275.0 1.003653544E9 5.582406057073806 0.7529332875079319 1.5722415400323753 0.3292090558828451 0.6910714863323157 0.02577555337930816
48630.0 1.061641624E9 1.9208786040692127 0.09031466602051914 1.3506065039940034 0.21694383601777456 0.7229121687906822 0.01876950586447982
49876.0 1.107410392E9 7.064506657257637 0.1708202860122061 1.7701408475473643 0.3574992319873054 0.7139310553653035 0.01432962989892049
51702.0 1.149064168E9 9.283586508053055 0.6414603799727803 1.12793541677748 0.18930458951378593 0.7143786177070958 0.008817034286516803
70348.0 1.135039096E9 3.932816792972239 0.6092641243325083 1.1693069734344674 0.4941098025556281 0.7049589509669353 0.008141281022841543
74978.0 1.23385896E9 2.426932483403231 0.31428682470647146 1.0870086081876171 0.20659435998579867 0.7204462063977267 0.016030465001562005
79201.0 1.30768132E9 6.418460635092642 0.5810773923204395 1.166485633857378 0.23746993715202766 0.7101477140228466 0.013473914170796316
85176.0 1.393157744E9 1.184470914832041 0.10107715281597253 1.7012496415722138 0.3995335283412699 0.7202398762791491 0.00782635395314254
89301.0 1.455334712E9 5.684078892397887 0.37615507533058334 1.6599569309530264 0.32343447313523055 0.7045865298504168 0.00758473619116184
96185.0 1.28324144E9 7.319507607384569 0.7315126605027208 1.0153267269058426 0.33660279363918477 0.7067246527605087 0.012074547614553706
97045.0 1.312284944E9 9.51802920569946 0.8976145120135676 1.1733254479345103 0.37414430487644434 0.7013951787827462 0.022094308373348394
102833.0 1.388376928E9 1.7050302288006303 0.23896274248266555 1.8167816553198142 0.20543990440885818 0.714600868873405 0.002085741291580042
111538.0 1.54661712E9 2.2130802871495914 0.7118060096652402 1.6840588243765502 0.46875187981425115 0.6950412227335856 0.009255470275353304
111868.0 1.568900064E9 4.894496901187777 0.3495421779824898 1.1738950460972846 0.26848741803479537 0.7191990726588201 0.022097908298988347
121557.0 1.67546256E9 5.3924435300248375 0.2425121107196942 1.197427457321235 0.3500783570894842 0.7130529587327702 0.009580636347955454
140489.0 1.556543688E9 7.461114328990838 0.8596664754043832 1.392347134012234 0.18152722239373673 0.6806821611851192 0.028670948315808756
142974.0 1.597683736E9 5.218303991020648 0.8676367182533121 1.972198391057401 0.4303996662608066 0.6721368425662538 0.004328388452869105
147476.0 1.651625624E9 8.036174047387362 0.9129158386603619 1.4308164824476644 0.30448794779088084 0.6984339989051289 0.013353620591657922
147730.0 1.675805136E9 8.921163654249808 0.9656971378021245 1.8116543957805584 0.22724217242201505 0.6636951907581815 0.020795025400285255
147758.0 1.679714752E9 9.668164313424306 0.3773293925592099 1.0151665689728122 0.3714132799212917 0.7190176321148612 0.024169711971744317
160139.0 1.629361432E9 7.832584209011712 0.7386643629359521 1.886457924309644 0.3850843762594095 0.6734634285433588 0.010917629983447787
166021.0 1.731287144E9 1.3811333366368361 0.014007220797994803 1.4872379652635046 0.44468594015751073 0.7167327007646368 0.007954168872637885
174154.0 1.826358744E9 3.2951994655398753 0.5892307812642472 1.54209476170346 0.29614317152020597 0.6981369710146392 0.03402833291333147
182754.0 1.923968776E9 6.903396779102156 0.1643050001335695 1.8253275495392594 0.41991146152179804 0.6957799584483154 0.01324647924891981
192011.0 2.02175304E9 7.311182151072626 0.6484917870505055 1.1541328026339568 0.3515970330945857 0.711583038325949 0.016786058311814317
196315.0 1.591508056E9 1.18108729949584 0.401175735131441 1.891891649708826 0.18575965653193607 0.699700802203005 0.02084890869398228
196327.0 1.592594256E9 3.8873688436245812 0.8853132770513075 1.7738159372028142 0.4154256114342184 0.682629195522328 0.013266562618951443
197200.0 1.634690568E9 3.5164899139662618 0.8819520570878033 1.0081140406912281 0.46798518275473466 0.7144341883736968 0.01000728449201773
213771.0 1.80994304E9 4.177110368223158 0.5532590696011848 1.3044098818263632 0.1903004517464491 0.709618363935855 0.008605632623559534
218831.0 1.888284272E9 3.545512855013082 0.8109125560396426 1.5096071651664982 0.49208668204810924 0.7036219519586117 0.009395901057234667
222934.0 1.943067952E9 1.5400919631098553 0.2829893108688261 1.8875484294834917 0.2246321064521577 0.7157171107198983 0.010734655249126607
222962.0 1.961852728E9 3.658174234786946 0.8737568241623955 1.4285382028387767 0.4223618805806696 0.6910746588418273 0.015851581204424193
242794.0 1.758368112E9 7.476376476754874 0.337850738548379 1.4296781364482793 0.24719650227858173 0.713201364677691 0.010038205948386988
245000.0 1.800679232E9 9.370017165607399 0.11375141470351133 1.646407291192437 0.4494709457695123 0.7109480999224035 0.005585327238154533
246665.0 1.840574264E9 5.095671196349531 0.8573374248604849 1.4014226073723013 0.16390018697373193 0.7014954930834879 0.010124099344752686
251227.0 1.918151568E9 2.7362735765907873 0.9864136855888085 1.934624960686064 0.13676471633425225 0.6701002264604788 0.014749887013671336
254185.0 1.96666384E9 5582.406057073806 0.7529332875079319 1.5722415400323753 0.061572037573640145 0.5523663599943472 0.04334687764130416
254907.0 1.971291784E9 9.96497724600189 0.05612611640414672 1.0834359529437907 0.3775511447043972 0.7187511593159757 0.011292863665187895
257626.0 2.024475656E9 5.230740186565407 0.5348256101060016 1.3608517832054585 0.14222789839826291 0.7042947219755767 0.019037808131940778
265508.0 1.778088664E9 342.67737363115833 0.5404634943325216 2.1836355075233818 0.05925004690535977 0.6731124994912739 0.015270509095333508
266428.0 1.847025616E9 1.9265000113381991 0.7109030215407427 1.0100736032378945 0.4178060916873182 0.7202839044111452 0.029770194768506685
274282.0 2.09867532E9 4.83593477775484 0.15092976629608099 1.8166258931440056 0.17387178321961516 0.6932631864030326 0.022483365626161846
281249.0 1.881700808E9 2.7690743136618345 0.16093375506486185 1.4936335124425941 0.39142865498441304 0.7187737538837147 0.01994430863967215
286713.0 1.962305704E9 7.457223263431855 0.5079668568466053 1.6215603768818934 0.36228152645183365 0.6958150540535323 0.010488156784317809
287292.0 1.988633528E9 5.16197868314025 0.615553759000578 1.7563421938726185 0.4619437095933707 0.6964531514665481 0.009603617573282753
292080.0 2.079647512E9 1.527550010761295 0.740697166353972 1.7572680848516566 0.15145212913517972 0.6993034375117052 0.0035446504861621287
293727.0 2.115503704E9 4045.2211076903745 0.5418650318132358 1.15506380864328 0.06410260332452332 0.6545033417010183 0.014878341928942334
294563.0 2.125494024E9 0.25163305858933893 0.46481190839252046 1.4824922550707125 0.23535841301416696 0.7146254884472142 0.014922544564620558
295929.0 2.155508528E9 342.67737363115833 0.5404634943325216 2.1836355075233818 0.05884254737441337 0.6791311415351625 0.023479206163966097
303978.0 2.19254576E9 2.4624364654236097 0.8174025041672938 1.1502518288632528 0.48315178280150217 0.7202227582278297 0.013892972909008264
309788.0 2.273454056E9 2.8176101971941323 0.4575477587989642 1.2706111274897673 0.48756478978463835 0.7160086260938613 0.00948390904823626
312984.0 2.332199376E9 0.7456447992776349 0.2293319245498495 1.946072463049354 0.24808400140905118 0.7108681841828165 0.025138593987299195
314633.0 2.395150192E9 0.6975301272663237 0.7233412664505098 1.7690119167781742 0.48888819197647293 0.6970509782588327 0.013302220806821701
321418.0 1.827417896E9 1.3983933694768902 0.9852447269108896 1.373261970300323 0.4809482219053224 0.7072075042080977 0.005279657152730951
336154.0 2.044452152E9 1.6129221361116786 0.3733492486128845 1.8146662978471406 0.47948339737603096 0.7104653564751388 0.014990730704236172
337283.0 2.090273456E9 3.946487058427236 0.6455145037568712 1.2360364909893051 0.27331149599557736 0.7065493502349519 0.006916545802158054
339001.0 2.142573664E9 9.807825867576703 0.4997269304832418 1.536639681141258 0.3935068022132546 0.6876287645110363 0.014920285261340066
344528.0 2.226778E9 9.60266936279813 0.20148109582311394 1.7203381737135124 0.14434597509853841 0.7045443422239384 0.0169090032230886
346156.0 1.992182984E9 4.310307142340647 0.6257558266468753 1.3574495400132478 0.25438387637753546 0.7089349288869169 0.011772806447186777
352430.0 2.081428968E9 9.69625966309308 0.7915715293919935 1.358117345181303 0.3021715550538516 0.7016607875790409 0.008716466946462657
359601.0 2.18156456E9 4.255873469053726 0.27917314862340803 1.4578043872410233 0.13596251569944112 0.7103817867245961 0.015357019219903305
372723.0 2.329972688E9 3.342269526738566 0.6111750065451231 1.932711633643239 0.16320727300878582 0.67952208470721 0.006370336454091284
375630.0 2.380136264E9 5.060490374759166 0.6215650630212081 1.0890105231170775 0.26643459744916653 0.7130898633434105 0.01218393543170073
377058.0 2.41718904E9 3932.816792972239 0.6092641243325083 1.1693069734344674 0.09867470557501631 0.6311985503332541 0.021493439832313017
377444.0 2.420830984E9 5.806818160478787 0.3269896604403513 1.68833845239295 0.24734524342181696 0.6976872621664119 0.022277061865239346
377621.0 2.441865968E9 8.448598169269488 0.6942477736664299 1.1916407096283863 0.49570193313348765 0.7070117491212239 0.008491218325435996
380378.0 1.907507864E9 3.232567307771034 0.7793842254229078 1.3387108418061757 0.35914051611781495 0.7051949856745671 0.029186244773557173
383273.0 1.970011016E9 9283.586508053055 0.6414603799727803 1.12793541677748 0.030093532640601835 0.6266777737799661 0.014159019286139609
394748.0 2.120197112E9 0.853428675890886 0.8572837137136211 1.5526826985039632 0.37040406162056605 0.7140870483329705 0.006018408064076753
399823.0 2.211202728E9 4.702935257088988 0.033372759131132336 1.3559126078812815 0.40114883458940465 0.7200895893285889 0.004671481426103248
400482.0 2.244187424E9 1.9761313419574766 0.8753151656860709 1.2154329768114471 0.46650318120424183 0.7078005339860008 0.00965804485999117
405247.0 2.314990136E9 4.525897385053258 0.2920216280669617 1.3938915071558535 0.10948404733341253 0.7163982372619152 0.021111461642853064
407655.0 1.751322328E9 2.9885031280753727 0.9222511938563279 1.399712533233474 0.11505680794167202 0.6940845738655526 0.016935980315308633
408323.0 1.785733376E9 7.305526860895975 0.10801087601452745 1.4216741120994143 0.3653346844884414 0.709359649160228 8.50804943391694E-4
421725.0 1.962826904E9 1.6925395384622555 0.6191798890940724 1.752683967602758 0.14889512826712054 0.7053755802159897 0.013890994892359291
424102.0 2.009328584E9 2.4298339806564098 0.054526993042224574 1.150696370233863 0.2055976683669084 0.7161843066205512 0.021511316220814102
431448.0 2.12271244E9 2.154536749894007 0.6218888672607388 1.7901065041555844 0.37366174019807596 0.6959460719463237 0.01296096699843944
432555.0 2.153152856E9 8.306765401809443 0.15656175706506584 1.5279887396150287 0.12678361032661534 0.7070825973336267 0.016547043259155958
433040.0 2.179154112E9 2.122159379548041 0.33486812109180686 1.9477707349916136 0.49175469610703126 0.7053852822450759 0.013165051552203808
440100.0 1.719810032E9 2.079286696675662 0.51528437943111 1.0247871499486831 0.282445124913424 0.7170362401746481 0.013079403901668429
449624.0 1.846824488E9 1.8160058667085255 0.8698515235624599 1.813832074007067 0.3626159732126578 0.6840760218599137 0.010435526027455847
451930.0 1.918044296E9 3.4882058150737736 0.8573183928123029 1.3694413401986822 0.35258212547938783 0.7164367843774798 0.015309577016638485
459395.0 2.02357656E9 9.60598813376839 0.4973381957023859 1.3518319616137893 0.4429055177995379 0.702068556301532 0.014646038577217628
462531.0 2.084370848E9 2.8854163010372647 0.5273533950063987 1.8427074452916645 0.2301383074325628 0.6981063934229672 0.0073419498452870135
466206.0 1.714297144E9 5.63834541739579 0.6309152332214293 1.7915423209918009 0.3945169540043306 0.6854168953798525 0.019689276388678248
471500.0 1.783692008E9 7.120682134134916 0.821495110611903 1.7534965088018686 0.2481733332983611 0.685361716214425 0.02028648328812727
472914.0 1.823612632E9 5.095323566083916 0.5434382920277592 1.582081107697701 0.2291941519902801 0.6998818062456552 0.00155788059716332
480099.0 1.946979416E9 0.9560949987965539 0.8603724930474578 1.1554803867046588 0.14032972839457747 0.7137559463403275 0.005583892630795802
485044.0 2.01049736E9 9.432034292532897 0.5000397934984788 1.7533728914650821 0.2700997659130243 0.6996255934775292 0.017244720425215472
486811.0 2.0540178E9 1.1789338668326055 0.02406665784648783 1.4349831423814168 0.162011900096125 0.7137534083327184 0.01190104568148355
495694.0 1.806886328E9 5.679768000398402 0.7662843674178953 1.3876986614125828 0.3083786697353899 0.6973619831912256 0.0056699628579325225
Altair EmployeeHi,
I forgot to mention this, but this has been fixed in RM Studio 6.0.005
Regards,
Marco
I forgot to mention this, but this has been fixed in RM Studio 6.0.005
Regards,
Marco