For our current release, we’ve introduced several exciting new features including: data connectors, visualization options, and the ability to fine-tune NN sentiment scores. Behind the scenes, we’ve made adjustments so taxonomies process faster and are easier to preview and filter. Read below to learn more.
New Features
AI Models
- The Tunable Neural Network Sentiment Model allows you to use a custom sentiment word list with our existing Neural Network Sentiment Model.
- Our Segmentation Model produces a segmentation analysis on structured data.
Data Connectors
- PubMed
- The PubMed data connector retrieves biomedicine and health sciences data from the U.S. National Library of Medicine.
- ClinicalTrials.gov
- Use this data connector to ingest all clinical trial information for a particular condition or set of conditions.
- Uploading a PDF
- You can upload a PDF, or multiple PDFs, to scrape the text for analysis, break the text into sentences and then index those sentences into a semi-structured data frame at the sentence level.
- Database connectors are now available for Snowflake, MongoDB, MySQL, Redshift and SQL Server. Contact the Stratifyd Solutions team for assistance configuring these.
Visualization Options
- Use a butterfly chart to visually display sentiment distribution within a given taxonomy topic or category. It's easy to read at a glance.
-
- This option is great for visualizing taxonomy labels.
- This option is great for visualizing taxonomy labels.
Improvements and Enhancements
- Significant changes were made to the taxonomy preview experience. Now, you can see how your taxonomy matches up against a dataset, but also work with a dataset next to your taxonomy as you are adding logic to a label. As you are editing logic in a label, your taxonomy preview will reprocess on the selected sample and point out the impact of your adjustments. This faster and can handle more records for a better representation of your data stream.
- We've updated the taxonomy filter menu to be more intuitive.
- You can use values created via aggregated calculations in visualizations by customizing the Y-axis.
- Part of Speech tagging allows you to see and understand the frequency of some basic linguistic units, enhancing information extraction.
- Chunk parser is now available as a standalone analysis.
We'd love to hear what you think! Feel free to reach out at any time to share feedback or suggestions.
Comments
0 comments
Article is closed for comments.