This project aims to explore and compare the effectiveness of machine learning models in stock market prediction across different stock types and under varying market conditions. This exploration is achieved by implementing and comparing the predictive accuracy of four Machine Learning models – SVR, XGBoost, LSTM and RNN – to understand how well they perform across diverse scenarios and stock profiles. These models will then be evaluated using XAI techniques, to hopefully identify the most influential model features driving predictions. Whilst there is an extensive volume of literature exploring and comparing a wide variety of models in their ability to predict the stock market, there is a limited amount of data exploring machine learning models with a specific focus on stock types and market conditions. This project addresses that gap by providing insights into which models are better suited for different stock profiles, and which models perform better in times of market volatility. The findings from this project aim to help practitioners and researchers in making informed decisions on model selection based on market temperature and stock characteristics