Does STATISTICA Data Miner support competitive evaluation of models?
We provide a wide selection of predefined Data Miner projects (or “workspaces”) that offer “competitive evaluation of models.” These projects can be modified, saved into files and later be executed either: (a) interactively – by clicking on specific analytic objects (represented by icons in the workspace) and dragging the connection arrows to modify the model, add filtering nodes, etc., or (b) automatically – e.g., by assigning the entire deployable project to a single toolbar button (with the entire workspace made invisible). For example, suppose you are trying to solve (interactively) a classification problem: 1.You would load the generic Classification Project, connect to it your data set (by dragging an arrow to connect the respective icons), define the variables to be considered (i.e., select the predictor variables from your input source), and then run it. 2. As part of the execution of this project, STATISTICA Data Miner will use all applicable classification algorithms (e.g., f
