Hybrid Machine Learning Technique for Personality Classification from Online Text using HEXACO Model
Personality refers to a person's unique collection of traits that influence their habits, behaviors, attitudes, and thinking patterns. Text accessible on social networking sites may be used to automatically identify an individual's personality characteristics. In the trials, a publicly acc...
Saved in:
Published in | 2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) pp. 253 - 259 |
---|---|
Main Authors | , , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
07.04.2022
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Personality refers to a person's unique collection of traits that influence their habits, behaviors, attitudes, and thinking patterns. Text accessible on social networking sites may be used to automatically identify an individual's personality characteristics. In the trials, a publicly accessible benchmark dataset from Kaggle is utilized. The major problem with the previous study was the skewness of the dataset, which was reduced by using the Re-sampling method, essentially random over-sampling, which resulted in improved performance. Although the results produced by all classifiers across all personality characteristics are acceptable, the performance of the XGBoost classifier is exceptional, reaching more than 99 percent precision and accuracy for various qualities. Individual preferences, talents, social and human values, personality variations, and human will power must all be addressed while analyzing entrepreneurial activities, intents, and performance, especially with entrepreneurship playing such an important part in the contemporary dynamic. Using the HEXACO Personality Characteristics Model, this study demonstrates a link between personality traits and entrepreneurial success, indicating that personality traits have a significant and direct effect on entrepreneurial performance. |
---|---|
DOI: | 10.1109/ICSCDS53736.2022.9760970 |