Automatic Annotation
This feature leverages semantic services to automatically annotate your data. You can apply automatic annotation at two different levels:
- Full table: Processes the entire table using a specific service like Alligator (Semantic Table Interpretation). This is currently the primary method for full-scale interpretation.
- Schema: Focuses exclusively on the table columns using a specific service, such as the Column Classifier.
Once the process is complete, the table updates automatically. For Full table annotation, cells are populated with predicted entities and metadata; for Schema annotation, columns are updated with Named Entity Recognition (NER) and kind classifications.
Semantic Table Interpretation
Click on the Automatic Annotation button in the Toolbar. Then, select Full table as target and Semantic Table
Interpretation (Alligator) as the method.
info
This is a long-running asynchronous process, allowing you to continue working on other tables while it runs.
Once the task is completed, you will receive a notification via a pop-up in the bottom-left corner of the screen.
Schema Annotation
Simply click on the Automatic Annotation button in the Toolbar. Then, choose Schema as target and Column Classifier
as method. This process automatically identifies the Kind of the column (e.g., whether it contains entities or literal
values) and assigns a NER Classification based on the cell values:
- For entities: It distinguishes between PERSON, LOCATION, ORGANIZATION, or OTHER;
- For literals: It distinguishes between NUMBER, DATE, or STRING;
This classification is crucial as it guides the system in suggesting only the most relevant Wikidata properties to your data type.
Once the task is completed, you will receive a notification via a pop-up in the bottom-left corner of the screen.