Welcome to Healthsea ✨
Create better access to health with machine learning and natural language processing. This is the trained healthsea pipeline for analyzing user reviews to supplements by extracting their effects on health. This pipeline features a trained NER model and a custom Text Classification model with Clause Segmentation and Blinding capabilities.
Read more in the blog post and visit the healthsea repository for all training workflows, custom components and training data.
Feature | Description |
---|---|
Name | en_healthsea |
Version | 0.0.0 |
spaCy | >=3.2.0,<3.3.0 |
Default Pipeline | sentencizer , tok2vec , ner , benepar , segmentation , clausecat , aggregation |
Components | sentencizer , tok2vec , ner , benepar , segmentation , clausecat , aggregation |
Vectors | 684830 keys, 684830 unique vectors (300 dimensions) |
Sources | n/a |
License | MIT |
Author | Explosion |
Label Scheme
View label scheme (6 labels for 2 components)
Component | Labels |
---|---|
ner |
BENEFIT , CONDITION |
clausecat |
POSITIVE , NEUTRAL , NEGATIVE , ANAMNESIS |
Accuracy
Type | Score |
---|---|
ENTS_F |
80.34 |
ENTS_P |
80.77 |
ENTS_R |
79.92 |
CATS_SCORE |
74.87 |
CATS_MICRO_P |
82.17 |
CATS_MICRO_R |
80.85 |
CATS_MICRO_F |
81.51 |
CATS_MACRO_P |
78.01 |
CATS_MACRO_R |
72.41 |
CATS_MACRO_F |
74.87 |
CATS_MACRO_AUC |
92.76 |
CATS_LOSS |
297.22 |
- Downloads last month
- 24
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Space using explosion/en_healthsea 1
Evaluation results
- NER Precisionself-reported80.770
- NER Recallself-reported79.920
- NER F Scoreself-reported80.340