The]: E [Randomized Algorithms] A health care company has access to the data of very many people storing their ZIP code, age, height, weight, blood pressure, and so on in a very large matrix, where we assume that all data can he stored as numbers. Here, each rowr corresponds to the data of one person. They wish to use this data to test some hypotheses they have about this population. In order to do so they wish to create a smaller or sparser matrix representative of the original very large matrix, with the data in the smaller or sparser matrix being still interpretable. W'hich randomized algorithm described in [1] would you recommend them to you use and why'iI I'Ib structure your answer, answer the following questions in the given order. Clearly justify your choice. [a] Would you recommend to perform element-wise sampling, row or columnsampling and why? {b} 'Which probability distribution would you recommend to use and why? Use examples of matrices to underline your argumentation why one probability distribution might be better than the other. To clarify: In case you opted for elernentrwise sampling you should be comparing two probability distributions here; for row or columnsampling there are three probability distributions you need to compare