Financial Market Sentiment Analysis and Investment Strategy Formulation of Social Network Data using Epistemic Neural Networks

Abstract

Financial market sentiment analysis using social network data involves extracting and analysing relevant information from social media platforms to gauge the overall sentiment of investors and traders towards specific financial assets or markets. The task involves utilizing social network data to perform sentiment analysis on financial markets. This manuscript present the Epistemic Neural Networks (ENN) optimized with Elk Herd Optimizer (EHO) for financial market sentiment analysis and investment strategy formulation (FMSA-ENN-EHO). Initially data is taken from Stock Marketdataset. Afterward the data is fed to Multi-window Savitzky-Golay Filter (MWSGF) based pre-processing process. After that the pre-processed data is fed into Synchro Spline-Kernelled Chirplet Extracting Transform (SSCET) extracting attributes such as raw data, stock quotes then textual data. The output of SKCT is fed into Epistemic Neural Networks (ENN) is to predicts the probability value, adaptability, robustness and interpretability. The weight limits of the ENN are enhanced using Elk Herd Optimizer (EHO). The suggested technique is applied in Python and the efficiency of the suggested technique FMSA-ENN-EHO is assessed with the help of several presentations evaluating measures in terms of accuracy is 98%, F1-score is 95%, Mean absolute percentage Error (MAPE) is 0.05%, precision is 95% and the recall is 97%,while comparing other existing methods such as stock price forecast procedure with leading pointers by Convolution Neural Network (CNN)then Long Short Term Memory (SPI-CNN-LSTM), BP Procedure in Stock Price Design Classification and Forecast (SPC-BPNN) and Forecast of stock price way utilized by hybrid GA-XGBoost algorithm (PSD-GAXA) respectively.