INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
The attributes which characterize the person such as emotions, behaviour, mind and temperature define a personality of a person. The project aims to determine personality traits on the basis of Big Five Model and which is considered to be the multi-label classification problem. This project covered various machine learning algorithm and also deep learning techniques for training and testing of models. The steps for building the model are collection of data, pre-processing, feature extraction, splitting data, training, testing and implementing the model. The algorithm used in project are KNeighborsClassifier, Support Vector Machine, Gaussian Naïve Bayes, Long Short-Term Memory neural network. This model successfully classify the personality. The results indicate that LSTM performed best after that Gaussian Naïve Bayes, Support Vector Machine, KNeighborsClassifier. As per results, Gaussian Naïve Bayes perform well by achieving accuracy of 87% and precision of 82% before applying Long Short-Term Memory, However The performance of Support Vector Machine after applying Long Short-Term Memory is outstanding by achieving more than 83% precision and accuracy of 80% for different traits.
Keywords:
KNeighborsClassifier (KNN), Support Vector Machine (SVM), Gaussian Naïve Bayes, Doc2vec, Long Short-Term Memory (LSTM).
Cite Article:
"Personality Prediction Using Deep Learning", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.6, Issue 6, page no.168 - 172, June-2021, Available :http://www.ijsdr.org/papers/IJSDR2106025.pdf
Downloads:
000337067
Publication Details:
Published Paper ID: IJSDR2106025
Registration ID:193405
Published In: Volume 6 Issue 6, June-2021
DOI (Digital Object Identifier):
Page No: 168 - 172
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn