a little cleaner, but still need to choose better variable names. the…

… ones from the paper aren't going to cut it.

voka/lof.py

@@ -1,39 +1,39 @@
-
-# https://www.dbs.ifi.lmu.de/Publikationen/Papers/LOF.pdf
-
+'''
+  Module that contains the function that calculates a Local Outlier Factor.
+  https://www.dbs.ifi.lmu.de/Publikationen/Papers/LOF.pdf
+'''
 import numpy
 
-def _distance(v1, v2):
+def distance(vector1, vector2):
     '''
     Euclidean distance.
     '''
-    a1 = numpy.array(v1)
-    a2 = numpy.array(v2)
-    return numpy.linalg.norm(a1-a2)
-                                
-def _reach(p, k, o, D):
+    array1 = numpy.array(vector1)
+    array2 = numpy.array(vector2)
+    return numpy.linalg.norm(array1-array2)
+
+def reach(p, k, o, D):
 
     '''
     k-distance of object p is defined as the distance d(p,o) between
     p and an object o in D such that:
-        
-    i)  For at least k objects o' in D/{p} d(p,o') <= d(p,o)
-    ii) For at most k-1 objects o' in D/{p} d(p,o') < d(p,o)
+        i)  For at least k objects o' in D/{p} d(p,o') <= d(p,o)
+        ii) For at most k-1 objects o' in D/{p} d(p,o') < d(p,o)
     '''
     distances = list()
     for op in D:
-        d = _distance(p, op)
+        d = distance(p, op)
         if d > 0:
             distances.append(d)
     distances.sort()
     kdistance = max(distances[:k]) if distances[:k] else 0.
-    return max([kdistance, _distance(p, o)])
+    return max([kdistance, distance(p, o)])
 
-def _lrd(p, k, D):
+def local_reachability_density(p, k, D):
     '''
     Local Reachability Density
     '''
-    denominator = sum([_reach(p, k, op, D) for op in D])
+    denominator = sum([reach(p, k, op, D) for op in D])
     return len(D)/denominator if denominator else 0.
 
 def LOF(p, k, D):
@@ -45,8 +45,13 @@ def LOF(p, k, D):
     return the LocalOutlierFactor for point 'p' compared
     to collection of reference points in 'D', using k-distance 'k'.
     '''
-    ratios = [_lrd(op, k, D)/_lrd(p, k, D)
-              for op in D
-              if _lrd(p, k, D)]
-
-    return sum(ratios)/float(len(ratios)) if ratios else 0.
+    ratios = list()
+    for op in D:
+        numerator = local_reachability_density(op, k, D)
+        denominator = local_reachability_density(p, k, D)
+        if denominator:
+            ratios.append(numerator/denominator)
+
+    result = sum(ratios)/float(len(ratios)) if ratios else 0.
+
+    return result