Fastest way to rank items with multiple values and weightings

Question

I have a collection of key value pairs like this:

{ 
   \'key1\': [value1_1, value2_1, value3_1, ...], 
   \'key2\': [value1_2, value2_2, value3_2, ...],
   ...
         


        

Answer 1

Here's a normalization function, that will linearly transform your values into [0,1]

def normalize(val, ilow, ihigh, olow, ohigh):
    return ((val-ilow) * (ohigh-olow) / (ihigh - ilow)) + olow

Now, use normalize to compute a new dictionary with normalized values. Then, sort by the weighted sum:

def sort(d, weights, ranges):
    # ranges is a list of tuples containing the lower and upper bounds of the corresponding value

    newD = {k:[normalize(v,ilow, ihigh, 0, 1) for v,(ilow, ihigh) in zip(vals, ranges)] for k,val in d.iteritems()}  # d.items() in python3
    return sorted(newD, key=lambda k: sum(v*w for v,w in zip(newD[k], weights)))