Personality characteristic forecasting method based on voices

Abstract

The invention discloses a personality characteristic forecasting method based on voices. The method includes the following implementation steps that personality evaluation and determination are carried out on multiple reference determination persons to obtain multiterm personality characteristic factor evaluated values; voice fragments of the determination persons are collected, multiterm acoustics prosodic characteristics are extracted, and multiterm statistical characteristic values are extracted; a voice personality forecasting machine learning model is built, the multiterm personality characteristic factor evaluated values of all the reference determination persons and the multiterm statistical characteristic values are respectively input into the voice personality forecasting machine learning model for training; the voice fragments of a determination person are collected, acoustics prosodic characteristics and statistical characteristics are extracted and input into the voice personality forecasting machine learning model to obtain the multiterm personality characteristic factor evaluated values corresponding to all terms of acoustics prosodic characteristics, and all the personality characteristic factor evaluated values of all the characteristics are weighted and summed to obtain and output the multiterm personality characteristic factor evaluated value of the determination person. The personality characteristic forecasting method has the advantages of being rapid in forecasting process and objective and accurate in effect, and forecasting material is easy and convenient to collect.

Description

technical field [0001] The invention relates to the field of computer application technology, in particular to a voice-based personality feature prediction method. Background technique [0002] At present, personality prediction methods on the Internet generally adopt the form of answer sheets based on text. Although the personality prediction method of the text answer sheet has a wealth of research results, such as the Big Five personality test (Big Five), Cattell 16 factor personality test (Sixteen Personality Factor Questionnaire, 16PF) and so on. However, in this way, users need to spend a lot of time to answer questions. The time required for prediction depends on the number of questions and the tester’s answering speed, and there are many and lengthy prediction steps. Testers are prone to boredom and resistance. The accuracy of test results depends on Due to the degree of subjective cooperation of the testers, this method is not very suitable for the simple and conven...

AI Technical Summary

Problems solved by technology

Therefore, how to overcome the shortcomings of the Internet and mobile platforms for predicting personality characteristics that take a long time, the effect is easily affected by subjective factors, and the measurement data is difficult to obtain, etc., and provide users with an easy-to-use "fast-food" personality prediction approach has become an important task. technical issues to be resolved

Images