The Five Pillars of Text Analytics: Search, Filter, Cluster, Code, and Classify
DiscoverText is a “Swiss Army knife” for text. With one platform, DiscoverText, users can capture, filter, de-duplicate, cluster, search, human code, and machine-classify large numbers of small, unstructured units of text. Our approach mixes human and computer training in an elegant and powerful loop. We provide a framework for users who build their own specialized text analytic approach. Backed by more than ten years of NSF-funded basic research, DiscoverText shortens the time it takes to create adaptive, custom machine-classifiers. Central to our vision is a firm belief in the power of systematic adjudication (or validation) of both humans and machines. DiscoverText is rooted in the emerging science of human annotation. The platform depends on accurate and reliable human input. Measurement metrics underpin our text classification success. DiscoverText is an advanced, multi-facted power tool for text.
text analytics features of DiscoverText include:
Schedule repeat fetches of live feeds via API Classification via manual training and automation Filter by metadata and threshold classification Redact and annotate sensitive information Code documents with ease alone or in groups Attach memos to documents, datasets and archives Connect and work with peers via your browser Measure inter-rater reliability and validate results Generate high-level summary and detailed reports Remove duplicates to limit time and effort Cluster near-duplicates and highlight unique text Bucket your filtered documents and search results Discover top meta values and unexpected concepts Build topic models to automate your groupings Enjoy a cloud hosted application with no installation Share projects, re-use models, and update results Watch some of our latest videos Register and login using Facebook or LinkedIn