You are tasked with developing a machine learning model to predict house prices. While preparing your dataset, you realize that an important feature, 'distance from the closest school,' has a significant number of missing values and does not exhibit high variance. Considering that every instance (row) in your dataset is crucial for the model's performance, how should you address the issue of these missing values?