Classification System for Detecting Human and Answering Machine Responses in the NuxCaller System by Nuxway Technology – Directed Research Project

The project consists of developing a prototype system for #recognizing human responses to answering machines for the NuxCaller system, which is an outbound #callcenter developed by NuxwayTechnology. The objective of this project is to develop a system capable of generating a report with call classification, accurately identifying which calls were answered by human and which went to voice mail. Each specific objective is a stage in the project to obtain the complete system. The project methodology is quantitative-experimental, where at each stage the operation is tested using recordings of real calls and at the end a significant sample of recordings is used, which allowed the accuracy percentage of the system to be evaluated. The results of the system classification were evaluated according to the two classification technologies being proposed as a solution, where both technologies showed a classification accuracy percentage greater than 90%, demonstrating the viability of the system for future integration into the #NuxCaller system. Keywords: #API, #LLM, #PBXIP, recognition, Vosk. The full project is available at the following link:

Ximena Guzman Cespedes-Trabajo Dirigido.pdf

Download PDF • 5.04MB