Have you ever imagined if humans and machines can work together? Humans are good at some things, and machines are good at others. Visualize what they can do hand in hand. If human coders and machines can consistently work with each other, they will be able to learn faster and create more accurate results. Thanks to Texifter LLC, DiscoverText makes this visualization possible.
What is DiscoverText?
DiscoverText is a cloud-based, collaborative text analytics solution by Texifter. It includes dozens of powerful text mining features and tools to help researchers make decisions more quickly and accurately. It is a platform that collects, cleans and analyzes text and social data, reducing the amount of time it takes for researchers to evaluate information. In a word, it is a comprehensive software solution.
It is an expert in text analytics, qualitative research, statistics, linguistics, and computer science. It combines adaptive and flexible software algorithms with human-based coding to conduct exact and consistent large-scale analyzes. It can merge data and pull text from various sources and combine information from multiple channels. Via Texifter’s Active Learning method, the software reduces the time it takes to make custom machine classifiers.
DiscoverText believes in the power of systemic validation of both humans and machines. As an advanced, multi-faceted tool for text, it combines human and machine training in a powerful loop. Its main features include clustering and filtering of information, and searching using keywords.
- Cluster. The software clusters information to help users find trends more effectively. It clusters data by the similarity of items to give users a high sense of social data landscape. By removing duplicates, clustering near duplicates and highlighting unique text, it helps them save time and effort.
- Filter. It filters in multiple ways. DiscoverText screens by date, metadata and threshold classification to create a more precise search. It stacks multiple filters such as complex search queries, machine qualifications scores, human coding and clustering for highly-refined results.
- Search. It allows users to search through raw data with the use of keywords. It allows them to discover top metadata values and train machine-learning classifiers to identify social media data and other relevant text.
How does it help researchers and companies?
Considered as a freemium “Swiss Army Knife” for text, DiscoverText has many valuable features that will help researchers and companies. The software allows users to manually and automatically classify data, filter information, and code documents easily. Relying on accurate and reliable human input, it allows users to connect and work with their peers via their own browsers. It also removes duplicates automatically, cutting days into hours.
Rooted in the emerging science of human annotation, the software also allows users to attach memos to archives, documents, and data sets. It also redacts and annotates sensitive data. It also allows users to schedule repeat fetches of their live feed via application program interface (API). Moreover, it generates high-level reports and measure inter-rater reliability by validating results.
DiscoverText also provides a framework for users to build text alone or in groups using their own text analytics approach. It builds topic models to automate groupings, clusters near duplicates, and highlights unique text. It also allows users to discover top meta values, share their projects, re-use models and update their results. In a few words, it allows users to search, capture, filter, cluster, de-duplicate, human-code and machine-classify large amounts of unstructured data.