Hybrid optimization enabled deep learning and spark architecture using big data analytics for stock market forecasting

Abstract

Summary The precise forecasting of stock prices is not possible because of the complexity and uncertainty of stock. The effectual model is needed for the triumphant assessment of upcoming stock prices for several companies. Here, an optimized deep model is utilized to effectively predict the stock market using the spark framework. Here, the data partitioning is done using deep embedded clustering, wherein the tuning of parameters is done using the proposed Jaya Anti Coronavirus Optimization (JACO) algorithm in the master node. The proposed JACO is developed by combining Jaya Algorithm and Anti‐Coronavirus Optimization algorithm. Then, important technical indicators are mined from divided data in slave nodes. Here, the technical indicators are considered features for enhanced processing. Then, data augmentation is done to make data suitable for processing in the master node. At last, the prediction was done in the master node using deep long short‐term memory (Deep LSTM), and training is performed with the proposed JACO. The proposed JACO‐based Deep LSTM attains the smallest mean absolute error of 0.113, mean squared error of 0.095, and root mean squared error of 0.309.

Identification of soluble solid content and total acid content using real-time visual inspection system

“…The framework involves utilizing machine learning algorithms such as C4.5 [26], [27], KNN [28], [29], and SVM [30], [31] for classification. To assess the framework's performance, you've listed several evaluation metrics, including classification accuracy, sensitivity, specificity, false positive rate, and false negative rate.…”