A Review of Recommender System Algorithms in Social Networks and Their Challenges

Today, the stock market has become a crucial channel for mobilizing investors’ capital. As a key indicator of a nation’s economic and financial activities, the stock exchange plays a pivotal role in reflecting the overall economic performance of a country or region. Predicting stock price movements remains one of the most challenging tasks in the financial domain. Accurate stock prediction not only enhances investors’ profitability but also contributes to national economic growth. Given the dynamic, complex, nonlinear, and nonparametric nature of stock markets, precise forecasting of stock price variations is essential for developing effective trading strategies. Researchers have employed various methodologies for stock market prediction, among which feature extraction and classification constitute the two fundamental processes. This study reviews and analyzes different feature extraction methods categorized into four types and classification techniques applied in prior research using artificial intelligence and mathematical models. The findings indicate that, due to the nonlinear nature of financial data, neural networks, particularly those employing hybrid or ensemble feature extraction approaches, demonstrate the highest efficiency and predictive performance in stock market forecasting.

کلیدواژه ها:

Recommender systems ، Social Networks ، Recommendation Algorithms ، Cold Start Problem ، User Interests ، personalized recommendations