...
- Original:
See all products in the order of your data feed. That means that the first product of your data file will be the first product to annotate and so on. - Random:
The product sequence you will see during the annotation process is completely random. That means that it is unpredictable which product will be presented next. Active Learning:
The product sequence is determined by the prediction model, prioritising prioritizing more informative products to be labeled first. The model will try to select products for which it is most uncertain what the correct attribute value is or which are most dissimilar from the already annotated ones with respect to the training attributes.Note title Caution This strategy is only appropriate when the selected training attributes are informative enough to determine the correct attribute value and when the model is already performing decently. If, in contrary, the model is performing very poorly, it might introduce sampling bias and require even more products to be annotated in this manner as opposed to applying the other strategies to achieve a similar model improvement.
As a rule of thumb, you might want to achieve at least 70% accuracy before using this strategy to reduce the annotation effort required to take your model to the next level.
...
Lastly, you have to decide which products you want to annotate. This step is only relevant if you already have product annotations. The two three decisions are:
- Remaining:
You only see products which have not been annotated at all. - Already Annotated:
Counterpart of Remaining - see all product which already got an annotation. This can be useful for refinements or if you use the filter option (see section below). - All Again (Rerun):
All products independent of current annotation status are shown during the starting annotation. All existing annotations are deleted.
Set Product Filters
Clicking the Product Filters button opens a sub window in which you can define attributes and its values all products have to fulfill to get displayed during the annotation. Setting these attribute conditions works like the Rule Builder Modal Window known from the Conditional Preselection or Questionflow configuration.
Furthermore, you can define, that all product variants are skipped during the annotation process. That means that for example a bike with different colors is displayed just once in one color. All other variants gets the same annotation value automatically.
If all settings are done, the Confirm button needs to be clicked to save the filter settings. The button caption now is showing "Edit Filter Settings" which underlines that the settings were saved.