How do I implement pearson correlation using map reduce?
I thought about mapping each pair (x_i,y_i), but I am not sure about the reduce part..
Break down the problem into parts
r(x, y) = (a - b) / (c * d)
Calculate a, b, c, & d and then insert them into the formula above. In some cases, it may be helpful/efficient to break down the formula further. If you want to perform each step using map/reduce, you'll need at least 4 separate map/reduce jobs (more likely 6 or 7). Only 1, (a), requires both x & y.