There is an error in Problem Set 10. File ps10.py Class Cluster Function computeCentroid
# Corrected
def computeCentroid(self):
dim = self.points[0].dimensionality()
totVals = pylab.array([0.0]*dim)
totVals2 = pylab.array([0.0]*dim)
for p in self.points:
totVals += p.getAttrs()
totVals2 += p.getOriginalAttrs()
meanPoint = self.pointType('mean',
totVals2/float(len(self.points)),
totVals/float(len(self.points)))
return meanPoint

Question

There is an error in Problem Set 10. File ps10.py > Class Cluster > Function computeCentroid
    # Corrected
    def computeCentroid(self):
        dim = self.points[0].dimensionality()
        totVals = pylab.array([0.0]*dim)
        totVals2 = pylab.array([0.0]*dim)
        for p in self.points:
            totVals += p.getAttrs()
            totVals2 += p.getOriginalAttrs()
        meanPoint = self.pointType('mean',
totVals2/float(len(self.points)),
totVals/float(len(self.points)))
return meanPoint

anonymous · Answer

that is not a mistake.
the centroid point is the mean vector of a cluster - it has no need for unNormalized data. 
All the functions that make use of it (the centroid) are using Normalized data.
No need for your correction.