Kingland announced enhancements to the Kingland Text Analytics Platform Suite, incorporating new administration menu options with deeper analysis and metrics for unstructured data sources and events.
Improvements to Named Entity Recognition (NER) training increased the suite’s accuracy to identify organizations and people, which are intuitively highlighted within documents, saving hours of reading and analysis time for teams manually reviewing relevant documents for specific entities and their related events.
“Industry leaders have shared with us that they want to efficiently solve their data challenges around searching and extracting data from a variety of unstructured data sources for use in onboarding, KYC, underwriting, compliance, risk monitoring, and other data-intensive activities,” said Chief Technology Evangelist Matt Good.
Teams can spend thousands of collective hours manually reviewing agreements or SEC documents as an example. While the average person can read and comprehend a 100+ page document in several hours, text analytics solutions can read and process the same document in minutes.
“Most enterprises have hundreds of thousands, if not millions of documents used by dozens or hundreds of processes,” said Good. “Organizations want speed, accuracy and the comfort of knowing that they’re making business decisions based off of extracted data that provides context with their counterparties, people, events and general entities of interest.”
Kingland incorporated targeted retraining to NER models, increasing the accuracy of stock models from 60% up to nearly 90% in aggregate over numerous source document sets.
The Kingland Text Analytics Platform Suite includes features such as:
Unstructured source integration
The nature of unstructured data – often referred to as dark data – makes it difficult for organizations to leverage business and decision making processes. Unlocking, or shining light on, data packed into unstructured sources enables organizations to make decisions based on more complete information. The suite integrates directly with RSS feeds, crawls bot friendly websites, and supports imported documents in PDF, HTML and other source formats.
Data identification & extraction
A combination of trained and configurable language models identify, tag and extract entities, people, events and other data attributes. Events in particular can be configured to specifically tailor unstructured data processing to focus on the insights required by client use cases and business processes.
Language modeling & training
Different models can be applied to different types of unstructured data sources and documents, supporting unique, fine-tuned analysis across legal documents, financial documents, news articles and more.