Published Paper Details:

Because of the intricacy, unconventionality, and nonlinearity of monetary frameworks, it is challenging to anticipate stock costs precisely. The PCC-BLS structure, a stock expense deciding multi-pointer feature decision method considering the Pearson connection coefficient (PCC) and Broad Learning System (BLS), was presented in this brief. PCC chose the initial stock price, technical indicators, and financial indicators as input characteristics from a list of 35 options. Second, a BLS was trained with these filtered input characteristics and information features were quickly extracted. The proposed strategy's performance was evaluated using four stocks that are either listed on the Shanghai Stock Exchange or the Shenzhen Stock Exchange. In addition, the forecasting results were compared to those of the Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Broad Learning System (BLS), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Multi-layer Perceptron (MLP), and Support Vector Regression (SVR). The highest modelfitting capability was demonstrated by the proposed model, which outperformed all other algorithms in this brief.