Thank you Wendy! ... View more

An added complication is, at least in many areas of the United States, parents have been attempting to name children with "unique" names. Below is a selected list of girl's names from a recent 10 year period. Careful reading will show that many of these names are somewhat phonetically equivalent to relatively common names, Aeryka <=> Erica for example. Acacia,Chili,Indigodawn,Mem' Ree,Secret-Destiny Aeryka,Cinamin,Indyana,Memphis,Serendipity Alaska,Clarity,Infiniti,Mesa,Shasta Alastrionna,Clarixxa,Innocence,Miami,Shasta-Rain Allyvia,Cloketta,Integrity,Mishalyn,Sha'uri Alpine,Creedance,Isabow,Modiesty-Star,Shy Ambrosia,Crimson,Itali,Monet,Sicily Americus,Cymmetry,Ixzy,Montana,Silver Amnesty,Cynica,Izrayelle,Mysticque,Sincere Anakalia,Daiquiri,Jaazminh,Nature,Snow Angelic,Daytona,Jetta,Nautica,Sonrisa Aptisam,Dayzee,Jewleah,Navy,Soul Aquilla,Dazsha,Jorja,Nirvana,Sparrow Arbor,Denym,Jubilee,Normandie,Starlit Arlington,Diamonique,Juniper,Northstar,Sublym Ataree,Diligence,Jupiter,Noxx,Supernova August Star,Divinity,Jurnee,Nutaliay Harmoney,Surreal Auktober,Dorcas,Jynnjer,Octayvia,Symphony Auroaramackay,Draekli,Kahlua,Olive,Tacheranai Autumn Hunnie,Dream,Kalispell,Otila,Tanaquil Autym,Dublin,Kanyon,Oyuky,Tehyanatane Aveda,Dymand,Karizma,Pallas,Teighlor Beatriz,Eboneyrose,Kaskade,Pandora,Tennessee Beautifull,Ecstacy,Kazpyr,Payshence,Thyme Belphoebe,Eeleceya,Kezzi,Peaches,Tottie Berlyn,Elexious,Khlover,Pennilane Meadow,Tragen Berthaalicia,England,Kiffin,Pepper,Tricity Bicardi,Envy,Klowie,Perfect,Trulie Blayde,Eos,Kozmo,Persephone,Trynadee Blessin,Epiphany,Krickette,Phaedra,Tsunami Blyss,Essence,Kronic,Poet,Tuesday-Rain Boisen,Eternytie,Krymsun,Poppy,Tundra Braenwynne,Fall,Kwincee,Prairie,Tyranny Breeze,Fancee,Lala,Pranaleyadri,Ugonna Bristol,Fashion,Lavender,Pranathi,Uneike Brittanica,Fayble,Lectra,Promise,Utahnna Brixx,Fayte,Lexington,Qatira,Vegas Brizzbin,Fennel,Libbertie,Quietstorm,Velicydee Brookenzie,Flossianna,Libertyann,Quimby,Viktoriya Burgandee,Freedom,Licet,Rainger,Wajd Byainett,Goldie-Moon,Little Summers,Ravenbella,Wrandie Byrkli,Graceland,Lixy,Rebel-Ann,Wyntre Cabella,Gyzzelle,London,Remedy,Xerenity Cachet,Hadies,Lotus,Remmington,Yafa California,Hailo,Love,Remzije,Ynfiniti Calloway,Happy,Lux,Reverie,Yochabelle Calypso,Harlequin,Lybburtie,Rhuivnyin,Zepplyn Capreece,Heaven,Magnolia,Russia,Zipaya Cascade,Hella,Maitre,Saig,Zipporah Cassiopeia,Hermyanie,Mali,Saylor,Zoigh Catalina,Hero,Malybu,Sã-kõ-yã,Zuzu Cedar,Heziachiah,Manhattan,Saoirse, Cedee,Holland,Maplejo,Sapphire, Celestial,Honesty,Mavity,Sativalyn, Celtic,Icelynn,Mayte,Season, Charm,Immaculata,McCall,Seattle, ... View more

Sorry this is the code that's creating errors /**********************************/ /*INFORMATION VALUE CALCULATION*/ /*******************************/ DATA META_VARIABLESET;   SET EMWS2.META_VARIABLESET    (KEEP=NAME NEWROLE);   WHERE NEWROLE NE 'REJECTED'; RUN; proc sql noprint;     select distinct name into :all_vars separated by ' ' from META_VARIABLESET; select count(*) into :num_vars from META_VARIABLESET; quit; %put There are --&num_vars.-- variables, listed below...; %put --&all_vars.--; /********************************/ /*Information Value Calculation*/ /******************************/ /* Check levels */ %let tiers = 10; ods output nlevels=checkfreq; proc  freq data=TRAINING nlevels; tables &all_vars. /noprint; run ; ods output close; 　 data  usevars; set checkfreq; where nlevels >= &tiers.; row_num = _n_; length str_var $34; str_var = strip(cat("'", strip(tablevar), "'")); run ; %macro woe_prep; proc sql noprint; select count(*) into :max_iters from usevars; quit; %put &max_iters.; %do i = 1 %to &max_iters.; proc sql noprint; select tablevar, str_var into :a, 😛 from usevars where row_num = & i.; quit; %let this_var = &a.; %let str_var = &b.; %put &this_var.; %put &str_var.; data temp; length tablevar $32.; set TRAINING (keep = TARGET &this_var.); tablevar = & str_var.; run; proc rank data=temp out=temp groups=& tiers.; var & this_var.; ranks var_rank; run; proc sql; create table temp2 as select tablevar, var_rank, avg(& this_var.) as mean_origv, min(& this_var.) as min_origv, max(& this_var.) as max_origv, count(*) as custs, sum(case when target = 1 then 1 else 0 end) as good, sum(case when target = 0 then 1 else 0 end) as bad, sum(case when & this_var. = . then 1 else 0 end) as miss_val from temp group by 1, 2 order by 1, 2; quit; %if &i. = 1 %then %do; data outstack; set temp2; run; %end; %if &i. >= 2 %then %do; data outstack; set outstack temp2; run; %end; %end; %mend woe_prep; 　 %woe_prep; 　 proc sql; create table outstack_tots as select tablevar, sum(custs) as tot_custs, sum(good) as tot_good, sum(bad) as tot_bad, sum(miss_val) as tot_miss from outstack group by 1 order by 1; quit ; 　 proc sort data=outstack; by tablevar var_rank; run ; proc sort data=outstack_tots; by tablevar; run ; 　 data woe; merge outstack outstack_tots; by tablevar; dist_good = good / tot_good; dist_bad = bad / tot_bad; woe = log(dist_good / dist_bad) * 100; pre_iv = (dist_good - dist_bad) * log(dist_good / dist_bad); run ; 　 proc sql; create table iv as select tablevar, sum(pre_iv) as iv, sum(good) as tot_good, sum(bad) as tot_bad, sum(custs) as tot_custs, sum(miss_val) as tot_miss from woe group by 1 order by 2 desc; quit ; 　 　 　 proc npar1way data=MT.BNA_HC_BGRE_BUILD_BALANCED edf noprint; var &all_vars.; class target; output out=ks_out (keep= _var_ _D_ rename=(_var_=tablevar _D_=ks)); run ; proc sort data=ks_out; by tablevar; run ; proc sort data=iv; by tablevar; run ; 　 data &init._iv_&model._&vers.; merge iv (in=a) ks_out; by tablevar; if a = 1; run ; 　 proc sort data=&init._iv_&model._&vers.; by descending iv; run ... View more

Hey Kanyange, How is it going with your logistic models? You are talking about two very important things: seasonality and choosing a good partition data set. A few comments below. Seasonality is related with certain patterns in your target variable. For example let's say that there is always a spike of events right on Easter, 4th of July, Thanksgiving, and the week of Christmas. If you were to train a model that predicts the probability of a customer taking your offer in the next 6 months, you need to take into account those spikes for both your input variables and your prediction. Your model the way you are describing it uses a whole year for the training set, so all these spikes are getting summarized. But remember to account for seasonal peaks if you start building variables that compare tendencies in two periods of time. For example for the variable (Balance in Q4)/(Balance in Q3) consider if you want to exclude or readjust the Q4 Balance to account the week of Christmas. In a similar way, you can train a model that uses the 12 previous months to predict a behavior in the next 6 months. But you will have to use your business knowledge to determine if this model will underperform for certain periods of time with unusual customer behavior that has been observed yearly for a given season. Choosing the right partition is also very important. There is no right or wrong approach. Different approaches suit different business needs. For the one you describe, training with a data set for 2014 to predict an event in 2015 is OK as long as you are using a partition node to train your model. In other words your flow would look like this: 2014 data set -> Partition (70% Training and 30% validation) -> Several model nodes ->Model Comparison. If you want to use a 2015 data set as a testing set, it needs to include the same variables as the 2014 data set, and the target variable. Set the role as "test" and connect it to the Model Comparison node. I hope this helps, Miguel ... View more

Hi, I built several incremental response models using both SAS Enterprise Miner and SAS/STAT.  The paper that Miguel cited was very helpful to me when I used Enterprise Miner to build models. Also, I would recommend you to take the SAS training course in this topic.  The course covers more techniques using SAS/STAT. Hope this helps a bit. ... View more

Here is a good place to get started: http://support.sas.com/documentation/onlinedoc/stat/930/introreg.pdf ... View more

If you want pick up variables , Check proc glmselect . ... View more

I think a statistically right way would be, if you run the regression again - without var1 & var6 - and compare the likelihood functions. If they don't differ much, you don't need the 2 variables. (Likelihood-Ratio-Test: Subtract lambda=2*(likelihood with var1&var6 - likelihood without var1&var6) -> get value of the chi-squared distribution with 2 degrees of freedom -> if the result is almost zero (e.g. <0.01) the 2 variables should not be excluded.) ... View more

hey Kanyange, did you get what you needed? The Transform Variables node has a "maximum normal" transformation method for interval inputs. I have not used it much as I am more of a binning or quantile grouping kind of person. But it seems to me that is what you need. Find more about it under the reference help (press F1) for the Trasnform Variables node. I hope it helps, Miguel ... View more

Hi Kanyange, Not familiar with arbitration. How different is that from model management (http://www.sas.com/content/dam/SAS/en_us/doc/factsheet/sas-model-manager-102722.pdf). In your formula, what does Value stand for, the value of the target? or is it a 2 stage model? Thanks, Miguel ... View more

H Reeza, The binning is a bit manual and is based on median , something like this below... So the final model looks like the file attached..Obviously something is wrong with the model..What do you think is the problem looking at the lift..? And the final equation looks like that...the variables have been regrouped by Variable Selection Node, that's why they have a G Prefixe. Your help would be much appreciated.. Thank you Parameter DF Estimate Standard Error aldChi-Square Pr > ChiSq Exp(Est) Intercept 1 -7.2641 0.788 84.97 <.0001 0.001 G_bin_aip_mth_03 0 1 0.2125 0.0048 1963.51 <.0001 1.237 G_bin_aip_mth_03 1 1 -0.3749 0.00515 5297.46 <.0001 0.687 G_bin_aip_mth_03 2 1 -0.0866 0.00307 795.05 <.0001 0.917 G_bin_aip_mth_03 3 1 0.0814 0.00281 841.56 <.0001 1.085 G_bin_prodosales_pfe_543 0 1 -0.2261 0.00367 3796.88 <.0001 0.798 G_bin_prodosales_pfe_543 1 1 -0.0544 0.00454 143.35 <.0001 0.947 G_bin_prodosales_pfe_543 2 1 0.0712 0.00357 397.86 <.0001 1.074 G_bin_prodosales_pfe_543 3 1 0.1613 0.00356 2052.07 <.0001 1.175 G_bin_prodosales_pfe_543 4 1 0.1879 0.00328 3272.51 <.0001 1.207 G_bin_prodosales_frt_43 0 1 0.1862 0.00407 2088.31 <.0001 1.205 G_bin_prodosales_frt_43 1 1 0.2896 0.00391 5484.97 <.0001 1.336 G_bin_prodosales_frt_43 2 1 0.2691 0.00415 4201.51 <.0001 1.309 G_bin_prodosales_frt_43 3 1 0.1832 0.00339 2928.51 <.0001 1.201 G_bin_prodosales_frt_43 4 1 -0.0497 0.00429 134.47 <.0001 0.951 G_bin_prodosales_frt_43 5 1 -0.3143 0.00507 3839.66 <.0001 0.73 G_bin_trans_34_pen 0 1 -0.6339 0.00267 56480.32 <.0001 0.531 G_bin_trans_34_pen 1 1 0.1799 0.00309 3382.14 <.0001 1.197 G_bin_wtr_trans_54_3 0 1 -0.2524 0.0037 4643.1 <.0001 0.777 G_bin_wtr_trans_54_3 1 1 0.2173 0.00492 1948.88 <.0001 1.243 G_bin_units_bst_543 0 1 -0.2345 0.00215 11879.58 <.0001 0.791 G_bin_units_bst_543 1 1 0.1112 0.00243 2095.4 <.0001 1.118 G_bin_sales_change 0 1 -7.9832 0.788 102.63 <.0001 0 /*Binning to normalize the variable*/ data &init._base_&model._&vers.;    set &init._base_&model._&vers.;    if &in_name. = 0 or missing(&in_name.)= 1 then &out_name. = 0;    else if &in_name. < 0 then &out_name. = -1;    else if &in_name. < 0.2 * &med. then &out_name. = 1;    else if &in_name. < 0.6 * &med. then &out_name. = 2;    else if &in_name. < 0.9 * &med. then &out_name. = 3;    else if &in_name. < 1.3 * &med. then &out_name. = 4;    else if &in_name. < 2.0 * &med. then &out_name. = 5;    else if &in_name. < 2.9 * &med. then &out_name. = 6;    else if &in_name. < 3.9 * &med. then &out_name. = 7;    else if &in_name. < 5.0 * &med. then &out_name. = 8;    else if &in_name. < 6.2 * &med. then &out_name. = 9;    else &out_name. = 10;   run; ... View more

Thank you..so what do you think of Proc Reg to get the VIF? ... View more

try LOG it  or use Cox-Box transform formula . ... View more