Keywords: third-party services, IBM Watson, Google NLP, text enrichment, categorization, topic analysis, sentiment analysis, emotion recognition
Apart from proprietary and open-source natural language processing methods and models, Dcipher Analytics also supports third-party services. From the Annotate Through 3rd Parties operation, you can run your data through Google NLP and IBM Watson Natural Language Understanding. The enrichment options covered are sentiment analysis, emotion recognition, categorization, and topic analysis.
1. Open the operation configuration window
Select the text field that you want to apply a third-party text enrichment operation to and click the "Add operation" button at the top of the workspace.
Search for "Enrich Text through 3rd Parties" or find the operation under "Text enrichment" and click it.
2. Name the output field
Under "Output field name", type the name of the output field.
3. Select third-party service
Under "Service name", select either "Google NLP" or "IBM Watson" to access the NLP services they provide.
4. Specify the language
In the "Language" drop-down, select the language of your text.
5. Select text enrichment options
The text enrichment options available are:
Categories: Labels each text with the categories that best describe it and a score from 0 to 1 describing the match between the category and the text. Labels contain multiple levels in a hierarchy of categories, from broad to specific and separated by "/" (such as "/business and industrial/energy/renewable energy/biofuel").
Entities: Labels each text with the entities identified in the text, their type, number of occurrences, and sentiment score (describing the degree to which the text about the entity is positive or negative). Entity types include organizations, locations, persons, and a variety of more specific entity types such as health condition, movie, and sport.
Sentiment: Labels each text as positive, neutral, or negative and scores the text on a scale from -1 (strongly negative) to 1 (strongly positive).
Emotions: Scores each text on a scale from 0 to 1 along the dimensions of sadness, joy, anger, disgust, and fear, based on the degree to which these emotions are expressed in the text.
6. Calculate the cost
The third parties apply a fee based on text volume. This fee is calculated when clicking the "Estimate cost" button. To reduce the cost, consider using the Preprocessing Wizard to remove duplicates and filter out long texts; the Sample operation to create a smaller dataset with a sample of data, and semantic filtering to remove irrelevant text.
7. Run the operation
Click "Apply" to run the operation. The outputs are now inserted into the output field.