Mining Model Prediction.

Mining Model Prediction... what for?

Which Data Mining algorithm use Mining Model Prediction?

Every algorithm has to use Mining Model Prediction for a final goal of a Data Mining Project?

In a very general sense...

Mining Model Prediction is typically the part of the UI which allows querying a mining model via DMX. What the model does depends on the algorithm, this part of the UI allows authoring DMX queries specific to that model.

Now, a mining model is not required to make predictions to be useful. Examples;

- algorithms may be used just to explore data, not to make predictions. Think the dependency net in naive bayes or Decision trees

- algorithms may assign unpredictive labels (e.g Clustering)

- algorithms may actually transform data for use in a different model (think of a decision tree model using as input the cluster label from a different model -- the cluster modell effectively transforms the input for decision trees)

|||

In case of Associantion Rules what should I show to final user? The same answer that Mining Model View gives or it′s necessary to make a Mining Model Prediction to a final answer?

If yes, it′s necessary to make a Mining Model Prediction to a final answer, how I do that?

|||

In general prediction isn't always the final outcome of a data mining models. Some models are only used for descriptive analysis. All of the algorithms in Analysis Services can be used for both predictive and descriptive analysis so it really is up to the application you require for your business problem as to how you want to use this.

If you want to get a predictive result from a model, you should always do a prediction and not try to impute the results from the algorithm content. The way to do so is to use a DMX query with a prediction join. To provide further details, I would need more information on what you were trying to accomplish.

|||Thanks