A Novel SRDPO-Based Weighted Dilated RNN for Automatic Personality Trait Prediction from Social Media Content

In recent years, the prediction of personality traits has gained more attention due to its capability to understand the individual’s unique personality. This work proposes an adaptive deep learning approach for personality prediction on the basis of social media text. Initially, the needed social media text data is gathered from the available resources. Further, the relationship between the words is analyzed by developing the Double Attention-based Bidirectional Encoder Representations from Transformers (A2BERT) for better understanding. Subsequently, the resultant features are fed into the proposed model called Adaptive Weighted Dilated Recurrent Neural Network (AWDRNN). Here, the weight optimization is done using the Stochastic Random number Derived Pine cone Optimization (SRDPO) to achieve an accurate predicted outcome. Finally, the results obtained from the developed approach are compared with other classifiers to determine the performance of the model. From the result, it is revealed that the designed technique attains high performance rates.