Use cases

Dashboard use cases

This dashboard allows you to see all usecases that you have created. You can access it by clicking on:

_images/side_bar_dashboard.png _images/dashboard.png

You can create a new usecase by clicking on top right button or you can check all informations on previously done usecases in the table. This table is sortable and filterable by usecase name / usecase type. Also, names can be clicked and will redirect you on the detailed view of the selected usecase.

On each line of the table, you can see:

  • The usecase name

  • The creation date

  • The data type (among « tabular », « time series », « images »)

  • The training type (among « regression », « classification », « multiclassification », « object-detection »)

  • The actual score, the optimized metric and a star system allowing you to quickly know if a model is well performing

  • The number of trained models

  • The number of predictions done

  • The number of people you have share the use case with (none by default)

  • The status of the usecase (running, done, crashed)

  • Actions linked to the usecase:

    • Pause or resume a running usecase
    • Share a use case
    • Make predictions
    • Stop or delete a usecase

Star system

By default, when no model is available, 3 gray stars will be displayed (and 0 blue). As soon as at least 1 model is available, the number of stars may change (up to 3 blue stars will be displayed according to the current performance of the modelisation).

The detail of the computation is explained here:

METRIC 3 STARS 2 STARS 1 STAR
MSE [0 ; 0.01 * VAR[ [0.01 * VAR ; 0.1 * VAR[ [0.1 * VAR ; VAR[
RMSE [0 ; 0.1 * STD[ [0.1 * STD ; 0.3 * STD[ [0.3 * STD ; STD[
MAE [0 ; 0.1 * STD[ [0.1 * STD ; 0.3 * STD[ [0.3 * STD ; STD[
MAPE [0 ; 10[ [10 ; 30[ [30 ; 1[
RMSLE [0 ; 0.095[ [0.095 ; 0.262[ [0.262 ; 1[
RMSPE [0 ; 0.1[ [0.1 ; 0.3[ [0.3 ; 1[
SMAPE [0 ; 0.1[ [0.1 ; 0.3[ [0.3 ; 1[
MER [0 ; 0.1[ [0.1 ; 0.3[ [0.3 ; 1[
R2 ]0.9 ; 1] ]0.7 ; 0.9] ]0.5 ; 0.7]
AUC ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
LOGLOSS [0 ; 0.223[ [0.223 ; 0.693[ [0.693 ; +inf[
ERROR RATE [0 ; 0.125[ [0.125 ; 0.25[ [0.25 ; +inf[
mAP [0 ; 45[ [45 ; 60[ [60 ; 100[
AUCPR ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
ACCURACY ]0.875 ; 1] ]0.75 ; 0.875] ]0.5 ; 0.75]
F1 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
MCC ]0.9 ; 1] ]0.7 ; 0.9] ]0.5 ; 0.7]
GINI ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
F05 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
F2 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
F3 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
F4 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
MACRO F1 ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
MACRO AUC ]0.85 ; 1] ]0.65 ; 0.85] ]0.5 ; 0.65]
MACRO ACCURACY ]0.875 ; 1] ]0.75 ; 0.875] ]0.5 ; 0.75]
QUADRATIC KAPPA ]0.8 ; 1] ]0.6 ; 0.8] ]0.2 ; 0.6]
MAP @ k ]0.875 ; 1] ]0.75 ; 0.875] ]0.5 ; 0.75]
LIFT @ k ]1 + 7 * (1-k) ; 1] ]1 + 3 * (1-k) ; 1 + 7 * (1-k)] ]1 + 1 * (1-k) ; 1 + 3 * (1-k)]

VAR is the variance of the target feature STD is the standard deviation of the target feature

Delete a usecase

In case a usecase is finished, it is possible to delete it by clicking on « remove » action link.

This action is irreversible.