Accern AutoNLP Taxonomy Framework™

Built for Financial Services
96.5% Accuracy

Curated & Continuously Updated over 5+ Years

With a million rows of training data, over 100,000 labels including out-of-the-box 60,000 equities and 250+ events, continuously updated with augmented AI / ML – so you don’t have to.

Entities

  • 20,000 U.S. Equities
  • 30,000 International Equities
  • 65 Commodities
  • 80 Country Exchanges
  • 20 FOREX currencies
  • 1,000+ Crypto Companies, Exchanges and Cryptocurrencies
  • + Bring Your Own (BYO)
    Simply upload files to BYO more Equities and Entities

Themes

  • 250+ Themes
  • 26 ESG Issues
  • COVID-19
  • M&A
  • Macro-Economic
  • Rumors / Speculation….
  • Regulations
  • Competition
  • Credit Risk

96.5% Accuracy – Purpose built for Financial Services

Core Taxonomy Model

Extracts and links entities

Spam Detector Model

Removes false positives

Verification Model

Assesses context to confirm true positives

Better, Faster, Cheaper NER with Taxonomies

Name Entity Recognition (NER) with the highest accuracy and better operational efficiencies (faster and cheaper) than NER with just Deep Learning models.

NER with Taxonomies is faster and cheaper

  • No need for large volumes of training data – Just start by adding keywords.
  • Easy to customize, maintain – just continue to add and refine keywords.
  • Transparent. Keywords and rules are easy to see, modify and understand.
  • Faster and cheaper than Deep Learning based models.

NER with Deep Learning is difficult and costly

  • Requires a training data set that may not be available or easy to acquire
  • Deep Learning cloud compute infrastructure is very expensive to operate, takes a lot of time and money
  • Difficult to keep updated and continue to maintain
  • Not Transparent – not easy to know why model came up with certain results

Tune / BYO Taxonomy

Simply tune our catalog or upload keywords – we make it easy to customize with suggested keywords to optimize your list. In addition, Expansion Rules, Lucene Queries and Exclusions further accelerate the building process.

BYO / Upload Training Data

Need even higher levels of accuracy? Upload your training data so we can further increase model accuracy for you.

“Leading NoCode NLP Platform”

“AI Breakthroughs”

“Multi-Level, Multi-
Language NLP”

Accern NoCodeNLP Platform

Connect

Pre-assembled Data Sources

Categorize

Pre-built Taxonomies

Analyze

Pre-trained NLP Models

Deliver

Pre-integrated Dashboards

Latest Resources

White Paper

A Benchmark of Popular Sentiment
Analysis Models on Financial News

Blog

Building Vs. Buying an AI Solution 

Extract Fast and Easy Insights
from Unstructured Data