BookmarkSubscribeRSS Feed
devi001
Obsidian | Level 7

Hi All,

 

We are doing comparison of SAS E-Miner projects to SAS Viya VDMML pipeline. We found that there is no node-to-node mapping available. Please refer attached snip for reference taken from Migration status document 2023.09.

 

Can someone please tell us what will be alternative in Viya VDMML for these E-Miner nodes - Append, Data Partition, Principal Components etc (there are few more but listed down here limited nodes). If there is any mapping available? 

 

Thank in advance!!

3 REPLIES 3
JosvanderVelden
SAS Super FREQ
There is a webinar available on moving: https://communities.sas.com/t5/Ask-the-Expert/How-Do-I-Move-From-SAS-Enterprise-Miner-to-SAS-Viya-Q-... . I don' believe direct migration of diagrams to pipelines was discussed. Maybe @MelodieRush has some ideas to share.
sbxkoenk
SAS Super FREQ

Sorry, I do not open attachments.


There are not so many Enterprise Miner nodes.

Here they are by the way: 

     Getting Started with SAS® Enterprise Miner™ 15.3
     About Nodes

https://go.documentation.sas.com/doc/en/emgsj/15.3/n1cpd0rgpneqwqn16mfcxp4sbjsb.htm


Can't you put your list as part of the text?

 

There's no 1-to-1 mapping
, but most of the Enterprise Miner capabilities are also there in Model Studio VDMML.

  • Appending and Merging NO, Data Partitioning YES (but it's done in a different way).
  • Also sampling the training data and scoring a hold-out set are done in a different way. Not through a (separate) node!
  • SOM/Kohonen EM node is no longer there in Model Studio VDMML.
  • Partial Least Squares EM node is no longer there in Model Studio VDMML. (that's a pity!!)
  • I am not sure about the availability of a cutoff node in Model Studio VDMML.

If you are unsure about an Enterprise Miner node, please ask ... and I will tell you how the same can be done / achieved in Model Studio VDMML.

 

BR,

Koen

devi001
Obsidian | Level 7

Thank you very much for both of your comments!!

 

We are trying to create below node wise mapping of E-Miner nodes and its Viya alternative. Basis on your inputs, updated the same for Append, SOM/Kohonen and Partial Least Squares. 

Also could you please validate this list for nodes to which we found mapping in Viya and explain for some nodes there is not direct mapping available but can be done in another way in Viya, is it possible to share that information (procedure/steps) over here?

 

Please Note - In below list, some nodes are in Magenta color is because we don't find it's direct Viya mapping but feature wise we feel that it could be alternative (we might be wrong in mapping some of those)

 

E-Miner NodeViya node
AppendNot available
Data Partition 
File ImportScore Code Import
FilterFiltering
Input Data 
Merge 
Sample 
Association 
ClusterClustering
DMDB 
Graph Explorer 
Link Analysis 
Market Basket 
MultiPlot 
Path Analysis 
SOM/KohonenNot available
StatExploreData Exploration
Variable ClusteringVariable Clustering
Variable SelectionVariable Selection
Drop 
ImputeImputation
Interactive BinningTransformations
Principal Components 
ReplacementReplacement
Rules Builder 
Transform VariablesTransformations
AutoNeuralNeural Network
Decision TreeDecision Tree
Dmine RegressionLinear Regression 
DMNeural 
EnsembleEnsemble
Gradient BoostingGradient Boosting
LARS 
MBR 
Model Import 
Neural NetworkNeural Network
Partial Least SquaresNot available
RegressionLinear Regression, Logistic Regression
Rule Induction 
TwoStage 
Cutoff 
Decisions 
Model Comparison 
ScoreScore Data
Segment ProfileSegment Profile
Control Point 
End Groups 
Ext Demo 
Metadata 
Open Source IntegrationOpen Source Code
Register Model 
Reporter 
SAS CodeSAS Code
SAS Viya CodeOpen Source Code
Save DataSave Data (Data Mining & Machine Learning)
Score Code Export 
Start Groups 
HP BN ClassifierBayesian Network
HP Cluster 
HP Data Partition 
HP ExploreData Exploration
HP Forest 
HP GLMGLM
HP ImputeImputation
HP NeuralNeural Network
HP Principal ComponentFeature Extraction
HP RegressionLinear Regression, Logistic Regression
HP SVMFeature Extraction
HP Text MinerText Mining 
HP TransformTransformations
HP Tree 
HP Variable SelectionVariable Selection
Incremental Response 
Survival 
Text Cluster 
Text FilterFiltering
Text Import 
Text ParsingText Mining 
Text Profile 
Text Rule Builder 
Text TopicCategories
TS Correlation 
TS Data Preparation 
TS Decomposition 
TS Dimension Reduction 
TS Exponential Smoothing 
TS Similarity 

 

Thanks once again for your help!

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 3 replies
  • 1001 views
  • 0 likes
  • 3 in conversation