Forecasting Stock Index Movement using Tree Based Classifiers: Evidence from an Emerging Economy

The given research addresses the issue of predicting the KSE-100 index through decision tree or random forest machine learning algorithms in an emerging market. The research attempts to build a complex analysis framework by including both types of technical indicators, e.g., stochastic %K and fundamental variables, e.g., oil and gold prices. The findings assert that the random forest model was better than the decision tree since the former has a prediction rate of 76% odds to 73%. Technical indicators are shown to be more effective as compared to the fundamental but combining them only provides a slight improvement in the predictive ability. The findings emphasize the prospect of ensemble learning techniques in the detection of non-linear market trends and associating investment strategies and risk management practices. The paper also proposes that emerging markets due to their structural inefficiencies can provide even more predictability thus adding to the growing body of research on the use of machine learning in forecasting financial markets.