Interprete predictions

Once the prediction interpretation options have been completed and initialized there is the option to interpret individual predictions.

1 | Choose analysis to work on [ Analysis 1 ]

2 | Operation mode of ML interpreter module [ Compute from scratch ]

3 | Overwrite existing MLIdatamat files [ Do not overwrite ]

4 | Overwrite existing permutations file [ Do not overwrite ]

5 | Recompute prediction change estimates [ no ]

6 | Select CV2 partitions to operate on [ 10 selected ]

7 | Save pre-processing params and models to disk [ no ]

8 | Use saved pre-processing params and models [ no ]

1 | Choose analysis to work on

The user can select the analysis to interpret.

2 | Operation mode of ML interpreter module

Here, the user can change whether the analysis should be computed from scratch or any precomputed MLIdatamat files should be assembled.

3 | Overwrite existing MLIdatamat files

Here, the user can specify whether existing MLIdatamat files should be overwritten or not.

4 | Overwrite existing permutations file

Similarly, the user can define whether existing permutation files should be overwritten or not.

5 | Recompute prediction change estimates

Here, the user can specify whether prediction change estimates should be recomputed or not.

6 | Select CV2 partitions to operate on

This option allows the user to select the folds to which the interpretation of the prediction is to be applied (see box “Selecting CV folds in NeuroMiner” in Preprocess features).

7 | Save pre-processing params and models to disk

The user can specify whether pre-processing parameters and models should be saved to the disk. This is not recommended due to space limitations.

8 | Use saved pre-processing params and models

The user can specify to use saved pre-processing parameters and models, if available.