A REVIEW OF PREDICTIVE MODELS APPROACHES TO STOCK MARKET FORECASTING IN KAZAKHSTAN

Abstract

. These days, the financial world is driven more and more by data, so predictive analytics has become increasingly important for navigating modern markets. This is especially true for places like Kazakhstan, where the stock exchange is expanding rapidly and the country’s becoming more integrated into global finance. In this review, recent developments in machine learning models for stock market forecasting are examined, with particular attention to their potential applications in Kazakhstan. The most common models - such as widespread use of LSTM networks - the types of data utilized, and the main evaluation metrics reported in the literature are summarized. Most studies tend to focus on complex models and large amounts of high-frequency data. However, in markets like Kazakhstan, hybrid, multi-scale models that also incorporate local sentiment and fundamental company data may offer the best approach. Ultimately, the key challenges involve ensuring high data quality and model transparency, as well as adapting global technological advances to fit local market realities. Through these efforts, stronger financial systems can be established to support more informed decision-making.