What are outliers and how do they impact data analysis and model performance? What methods can be used to detect outliers in a dataset? How can techniques like removal, transformation, and capping be applied to handle outliers? When should outliers be retained instead of removed? How does the choice of handling outliers affect the accuracy and reliability of models?