Stock market prediction under a deep learning approach using Variational Autoencoder, and kernel extreme learning machine

Forecasting stock market trends is challenging due to its complex and dynamic nature. Stock market prediction is important because it helps investors make informed decisions and allows businesses to plan their financial strategies. Prior works introduce several Artificial Intelligence (AI) based works, however, they are not robust in semantical pattern identification from the dataset. In light of the above limitation of existing works, this paper introduces the VAE-KELM-based mechanism. In this work, the convolutional neural network (CNN) technique was implemented to extract features and a variational autoencoder (VAE) for predictions. The goal is to predict the stock price with high accuracy. The technical indicators and historical data were considered input, with the sub-maximum layer being substituted by the kernel-based Extreme Learning Machine (KELM). In addition, the application of Autoencoder (AE), Variational Auto Encoder (VAE), Extreme Learning Machines (ELM), and full writing from Radial Basis Functions (RBF) is necessary for demonstrating the multiple stock market sources. VAE has been used for future extraction over the extracted feature. ELM mechanism is used for prediction and also this research used historical data and technical indicators to predict stock market trends. Various performance metrics were used to evaluate the proposed model's accuracy, and the results were compared with benchmark methods. The study's results demonstrated the proposed algorithm's effectiveness and superiority of the proposed model compared to other methods.