It is definitely a good suggestion. I will forward it to in product management. Any more suggestions please send them our way. Thanks Jakes! ... View more

Hi Jake, Newer versions of Enterprise Miner have the option to use the decision tree node to import a tree model ... we could plan a workaround with EM13.1 or newer. For EM 6.1 I think the easiest might be to use a decision tree with the same settings of the proc arbor that runs in the rule induction node. From the log, I got: Decision Tree Node Option Value Leaf Size 50 Minimum Categorical Size 5 (default) Maximum Depth 4 Missing Values Use in Search (default) Nominal Target Criterion Gini Assessment measure Average Square Error I got a slightly better model, so I may be missing one more option. Hopefully someone will come to the rescue! In the meantime this should get you close enough. Good luck! Miguel ... View more

Hi Ken, I don't think the colors mean anything, they just help you know what to look and where to click. Or at least that is how I use them :smileysilly: For the example below, you know that Cluster node found you 6 segments or clusters. Segment 1 has around 60% of customers with low values of savings (between 1-1.5), around 20% of customers with high savings values, and 20% customers in between. Personally I just use the colors in the bar as a reference to get the mouse-over box (a nice summary), and the default theme has never got me confused I think... But for the most part I use the table that feeds this graph. I click back and forth a lot between the plot and the table when I really want to understand the distribution of a variable within a segment. An alternative is to use the Segment Profile node, but it is not as interactive... Does this answer your question? As a follow up, would you find it more useful if there was an option that you can choose to make these colors have a meaning, like for example a gradient legend? Thanks, Miguel ... View more

Hi Hussein, A great thing about Enterprise Miner is that you can try multiple techniques at once. Since two flows on a diagram run in parallel you are also making the most out of your time. Note that nodes like decision tree will do variable selection, and nodes like PCA will do dimension reduction. Advanced Predictive Modeling Using SAS Enterprise Miner is a course that explains very well advanced topics, including unsupervised (PCA, variable clustering, etc) and supervised (PLS, LARS, LASSO, etc) dimension reduction techniques. Highly recommended! A best practice is to try several techniques and select the one that suits your target, number of observations, and number of input variables. Also check this discussion ( ) where you can find a list of nodes that you can use for variable selection. Also, what kind of classification of problem are you trying to solve? Are you dealing with missing values in the input variables? And what is the distribution of those 6 classes? You might even want to use a two-step model. I hope it helps, Miguel ... View more

Hi Hussein, By default your nodes only run once, unless you change a property on that node. Even when you change properties, Enterprise Miner is very efficient about what code to re-run. For example, if you change a property in the "Report" part, only the code for that portion is re-run, saving you a lot of time. I think by default EM is doing what you want, even if you see the green status symbol of the nodes swirl for a quick fraction of a second. I hope it helps, Miguel ... View more

Hi Ken, Please provide a little more background. I frequently use decision trees to model each segment of a cluster node, but the target variable is not the segment variable. For example for the German Credit data set I can add a cluster node that will give me X segments. For each segment I add a decision tree that will model the probability of default (variable BAD). Variable bad is the target, not the variable segment created by the cluster node. Does that help? What are you trying to do? Thanks, Miguel ... View more