Get a glimpse into the core features and capabilities of PrediCX™.
Want to see more? See our pricing and sign up for our free 30 day trial.
The latest in machine learning for Voice of Customer data and text data, PrediCX™ essentially acts as a supercharged ‘labelling engine’ taking the minimal human input and replicating that domain knowledge to label concepts (labels are also known as “classes”). These can be multi-labels per record. For example if a record is a query from a banking customer about how to obtain a credit card statement, then the labels can be whatever you choose e.g. “query” (as opposed to complaint), “credit card statement” (subject matter) and maybe the emotional state of the user e.g. “neutral”. These can then be used to drive insight and/or automation. These labels all have explicit performance.
The system can be used via an API to feed labels and metadata back into your BI and analytics platforms. Alternatively it can be viewed in the GUI.
The performance of the labels (aka tags or models) can be inspected and improved in terms of metrics like F-score, precision, recall and accuracy.
The labels can be easily structured in hierarchies and moved or merged between nodes and with other labels to suit the analysis and business need.
The core technology is called Optimized Learning (“OL”) which in a nutshell makes machine learning simple to apply so that anyone can set up and run it, not just a data scientist. OL steers human users or curators to import any labels that might exist to help train the models, and then to label certain specific records where there is uncertainty in a highly efficient way to minimize the human input whilst maximizing the performance of the overall system. This improves and maintains accuracy of the labels overall as well as helping to spot new topics e.g. early warning.
PrediCX™ can be used as a standalone product, or it easily interfaces with other systems via RESTful API on-premise or in the cloud.
Classify by Concepts
Tunable Sentiment Analysis
Early Warning Capability
© 2017 Warwick Analytics. All rights reserved.
Registered in England & Wales. Number 07724630. Registered address 35 Kingsland Road, London, E2 8AA. VAT 120435168.