An Optimized Combinatorial Approach For Stock Market Prediction Considering User Preferences

The economy is often seen as the engine that propels a nation forward. It regulates the distribution of resources by keeping an eye on the complex relationship between production and consumption. To put it simply, the economic system is there to meet the demands of all the people and businesses that make up the economy. The stock market is an important part of any economic understanding since it is the setting in which activities such as the purchase and sale of shares and the issue of new shares by different types of limited liability companies (both public and private) take place in relation to market research and exchanges. Because it is so difficult for the average person to predict when stock prices will go up or down, it is critical to find the best stocks on the market and put their money into them. When it comes to intelligently and reliably meeting user demands, however, conventional processes fall short. An optimized stock prediction system that takes user preferences and demands into account is presented in this study as a new way to manage retrieval. This technique is crucial for properly satisfying user requirements. Essential aspects such as supply-demand dynamics, inflation, geopolitical events, technological advancements, natural catastrophes, company performance, and the efficiency of government data may be dynamically gathered using this technique. All of these things add together to make stock price swing predictions. Machine learning algorithms are crucial for stock prediction because of the massive amounts of data that must be collected, processed, cleaned, and analyzed. The proposed method, ML + LSTM, yields 96% accuracy. Comparing the existing methods, which are KNN, decision tree, and random forest, gives accuracy like 80, 92, and 94 percentages. In addition, the article highlights the use of ML techniques including k-nearest neighbour, decision tree, and random forest. Important financial indicators, including common factors, sector factors, and stock-specific variables, are improved by adding more elements to these algorithms. The ultimate aim is to keep everything in check when handling massive datasets in order to accurately forecast stock prices. Simulation experiments implemented verified the performance of the proposed system over the existing retrieval management scheme based on training and testing the accuracy level to invest on best stock.