Intelligent algorithm for automated processing of citizens' appeals – Tula, Sverdlovsk, Chelyabinsk regions, Republic of Tatarstan, Khanty-Mansi Autonomous Okrug
Automation of processing of unstructured data streams (citizens ' appeals) using machine learning and neural network modeling methods.
The problem of manual processing of requests and their distribution by topics and categories.
A large number of requests, some of them are typical and fall into the "Other" category, although in fact they relate to existing topics.
Citizens ' appeals are sent to various information resources of the regions, divided into categories and topics. Depending on the source from which the citizen's message came, or how an employee of the executive authority chose the topic, the appeals are grouped, but employees have to read the text of the appeal and look for a suitable template for processing or responding.
The service allows you to:
Determine the subject of the message with high accuracy, as well as the responsible performer.
Analyze the written request of the applicant, make stylistic recommendations for the preparation of the response in order to prevent repeated requests.
Reduce the time for processing citizens ' requests, improve the efficiency and quality of responses to the citizens of the region who applied.