BACKGROUND
A business information management software-based product stores and manages the operational data of the companies’ key processes.
Problem
A company has a tool that captures and classifies Consulting the business data managed by this product is usually complex since technical and programming knowledge is required; unusual skills in a typical end-user profile. For this reason, it is necessary to incorporate the possibility of querying the data in a less technical, more natural way.
Benefit
Through the development of a new module based on the use of NLP techniques, it is possible for users without a technical profile to consult their business data in an extremely simple way.
METHODOLOGY & results
Architecture: Cloud architecture based on AWS services and databases, using RedShift as a data warehouse.
Developing language: Java.
ML techniques: NLP techniques and unsupervised learning algorithms (clustering).
Results: The system incorporates automatic learning capabilities, capable of resolving ambiguities and improving its results through use, managing to resolve more than 90% of correctly formulated queries.
