Generating data features from speech and sentiment analysis for improved predictive routing

Abstract

A method for processing data to train a predictive routing model. The method includes receiving interaction data from previous interactions, including audio data capturing a conversation and a transcript of the conversation. The method then performs speech analysis by processing the audio data to determine speech metric scores, which include a measure of how much the agent or customer speaks during the conversation. The method continues by performing sentiment analysis to determine scores associated with sentiment metrics. Sentiment metrics include a measure of sentiment based on classifying statements in the transcript data as positive or negative. The method then performs feature engineering to generate feature data and create a training dataset from it.The method continues by applying a machine learning algorithm to the training dataset to train a predictive routing model.