# Dimensionality Reduction **Musing** * In some ways, reframing a problem or viewing it from another angle is a form of dimensionality reduction. We are taking a problem with many dimensions and view points and compressing it down into several. This can equivalently be thought of as using heuristics. The idea of framing is that context/background info (prior knowledge) is what allows us to pick to dimensions to utilize. We pick the dimensions based on their correlation to our making a good decision, and other factors (information content?) * Models allow you to see a problem from additional dimensions, where new dimensions yield new infortmation and context. Sometimes we need to drop old dimensions that aren’t useful (horse racing)