The easiest way to view the surrogate rules is to go into the Interactive tree utility once your model has been fit (unless you are already there).   To access this functionality, click on the desired Decision Tree node and then click on the ... to the right of Interactive.   Once the Interactive utility opens, click on the desired node and then click on View --> Surrogate Rules Detail to see the splitting rule(s) that were created.  Let me know if you have any trouble locating it.   Hope this helps! Doug ... View more

By "f-score", are you talking about the traditional F-measure or balanced F-score (F 1  score) which is the harmonic mean of precision and recall or are you referring to an F-test statistic of some sort?  SAS Enterprise Miner calculates precision defined as % true predicted events / (true predicted + false predicted) and recall defined as the event classification rate.  It is likely you could compute this statistic relatively easily but I am not aware of anything by that name being generated by SAS Enterprise Miner.  Do you have a formula for the f-score you wish to compute?   Cordially, Doug ... View more

From the SAS Enterprise Miner help, the File Import node is designed to convert selected external flat files, spreadsheets, and database tables into a format that Enterprise Miner recognizes as a data source and can use in data mining process flow diagrams.  The files you are describing are text files that can be used by text miner.  These files would not need to be imported into SAS Enterprise Miner to be used as text; instead, they would need to be put in a particular location with other text files where the Text Miner nodes could access them and analyze them.   I hope this helps! Doug   ... View more

SAS Enterprise Miner generates a ROC curve for the Train, Validate, and Test data set in the Model Comparison node when modeling a binary target.  It also generates a misclassification chart for the Train & Validate data sets but it does not generate a misclassification chart for the Test data set.  In the design of SAS Enterprise Miner, Test data sets are intended for a final unbiased evaluation of model performance so they are not used by default when a Validate data set is present.     Please note that SAS Enterprise Miner always generates       F_<target variable name> : the target variable value        I _<target variable name> : the predicted target value (based on highest probability)   but it can also generates a D _<target variable name>  which contains the 'decision' outcome based on the decision weights and priors entered in the target profile when one is present.     For example, if the target variable is named 'BAD', SAS Enterprise Miner would create the variables F_BAD, I_BAD, and D_BAD.       If a Test data set is available, you can add a SAS Code node after any modeling node and enter the following code in the Training code section.  This example assumes the target variable is named BAD.     /*** BEGIN SAS CODE ***/   proc freq data=&em_import_test; tables F_BAD*I_BAD; tables F_BAD*I_BAD;  *only available if Decision profile has been created; run;   /*** END SAS CODE ***/   The code above will generate both misclassification charts if the target profile is available.    Hope this helps! Doug ... View more

If you go to the SAS Enterprise Miner help available by opening SAS Enterprise Miner and then clicking on Help --> Contents, you can navigate to the Cutoff Node by navigating in the panel on the left to   Node Reference     Assess Nodes          Cutoff Node   and then navigate in the panel on the right to Cutoff Node Train Properties, you can scroll down until you see Event Precision Equal Recall where it says the following:  Event Precision Equal Recall — With precision defined as % true predicted events / (true predicted + false predicted) and recall defined as the event classification rate, this method chooses the point at which precision and recall are equal.   There are two ways to find this in the output:   (1) In the Overall Rates plot, a line is drawn at the requested point and hovering over the line with the mouse will show the cutoff  (see attached document for plot) (2) In the Output section, you can see the point at which the first two columns are closest is when the cutoff is at 0.36 (I added the bold -- see attached document for partial table).     -------------------------------------------------- |           |            |             |             |Overall | |           | Event  |  True   | False    |Classif-| |           |Precisi-|Positive|Positive|ication | |           |on Rate|  Rate    |  Rate   |   Rate  | |--------+--------+--------+---------+---------| |Cutoff |            |            |             |             | |--------|             |            |             |             | |0.99    |  200.00|    8.30 |    0.00 |  161.78| |--------+--------+--------+--------+----------| |0.98    |  200.00|   13.26|    0.00|  162.77 | |--------+--------+--------+--------+---------| |0.97    |  196.67|   15.47|    0.07|  163.15 | |--------+--------+--------+--------+----------|     .            .        .        .        .   .            .        .        .        .   .            .        .        .        . |--------+--------+---------+--------+---------| |0.38    |  134.11|  125.06|   15.31|  172.80| |--------+---------+--------+--------+---------| |0.37    |  132.51|  127.47|   16.18|  172.59| |--------+---------+--------+--------+---------| |0.36    |  131.01 | 129.67|  17.01 | 172.36| |--------+---------+--------+--------+---------| |0.35    |  128.45|  131.27|   18.22|  171.71| |--------+---------+--------+--------+---------|   I hope this helps! Doug ... View more

On the PROC HPSPLIT statement, there is a PLOTS option that will allow you to open up the subtree where you start and to a set depth.  In complex trees, you will not be able to reasonably see the entire tree in one plot without losing many details.  The code below refers to the SAMPSIO.HMEQ data set which is available as a sample data set in SAS Enterprise Miner and is also attached here. The code requests the displayed Tree to have a depth of 5 beginning from node "3":   /*** BEGIN SAS CODE ***/     libname x 'c:\data';   * < note: change libname and path as needed > ;   proc hpsplit data=x.hmeq seed=123 maxdepth=10 plots=(zoomedtree(nodes=("3") depth=5));    class bad reason job;    model bad (event='1') = debtinc derog loan mortdue value job reason clno ninq yoj;    grow entropy;    prune off;* costcomplexity(leaves=all); run;   /*** END SAS CODE ***/   I hope this helps! Doug   ... View more

IrinaN,   The short answer is that there is not a catalog for functions that are (generally) only used in SAS Enterprise Miner since these would typically provide no benefit to the user, but if you have questions about what a particular function does, you can look at the code (as you have done) or inquire with SAS Technical Support.      In this case, the DMNORM function you are mention is used for normalizing input field names and values to have no more than 32 characters in a the name and no more than 32 characters in the field. It uses the internal normalized version of the variable for analysis and in the score code it generates but you never would have need of these functions elsewhere.    The normalization in this situation is important because many data management applications/utilities export data with unnecessarily wide fields (e.g. 200+ characters for a Yes/No variable).  Since SAS Enterprise Miner is designed to generate score code and the entire potential width of the field must be stored just in case it is needed, this limit prevents the data from becoming unnecessarily large and it prevents the scorecode from becoming unnecessarily long as both of these will slow processing.   Even if your grouping variables have levels that do not differ prior to the first 32 characters, SAS Enterprise Miner will still keep them distinct but you will have to go back to the code in order to figure out which level each normalized level is assigned to.  This is why we recommend to make sure that you don't use unnecessarily long field names/values, but if you do then make sure they differ in the first 20-25 characters so they will be easily distinguished.  In general the DMNORM function handles all this but it is not a function that would typically be used directly by a user.      I hope this helps! Doug ... View more

It really doesn't matter to an algorithm where the data came from or whether or not there should be 'missing' values or not.   The data structure for the techniques you are describing anticipate that there are going to be distinct units/observations/entities on each row (not spread across multiple rows) and each column will contain an attribute for the unit/observation/entity on the corresponding row.  So if we were looking at cars, your rows might correspond to a particular make and model of a car and the columns might correspond to things like suggested retail price, city mpg, hiway mpg, number of cylinders, drivetrain type (front/rear/all-wheel), bluetooth enabled (yes/no), etc...  It is possible that you don't have complete information even in simple situations like this since Mazda doesn't have cylinders (its a chamber) in its rotary engine, and some models might not post certain information.   It is important to note that a neural network, a support vector machine, or a regression model will drop any observation with incomplete data which simply means there is a missing value for one or more of the input variables. Decision Tree models are able to incorporate these observations but you must impute/guess the missing value if you want the observation to be considered at all in your neural network or regression model.  Adding rows with incomplete data will not help these latter modeling types but even incomplete data can be used by a Decision Tree model.   If the rows that have been 'added' are not really contributing any additional information to the model, it is possible that one of those methods requiring complete data might be helpful.   From a method standpoint however, it is important to understand how the methods are interpreting your data and to decide what will generate meaningful result. I hope this helps, Doug  ... View more

Kamal, The SAMPSIO library in SAS Enterprise Miner actually consists of several sample library locations and I would not recommend putting your data in any of those.  Instead, you can simply copy the data to a location that is accessible from SAS Enterprise Miner and then click on File --> New --> Library... to define the path to the data location (relative to the server where SAS Enterprise Miner is installed).  Note that you cannot use locally mapped drive letters but must specify the fully qualified path (e.g. //myserver/myfolders/myprojects/mydata) for any drive which is not local to the server doing the work.   You can then define a new data source using the LIBNAME you defined in the wizard launched by requesting to create a new library.    Let me know if you have any trouble.   Cordially, Doug ... View more