+/OCR

character recognition demo

git clone https://git.julia.blue/+/OCR

Joseph RyanAdd weightings10501ef

master
1.2 KiB40 linesraw
1from collections import Counter
2
3
4def load(fileName):
5    f = open(fileName)
6    lines = [[int(i) for i in l.rstrip('\n').replace(" ", "").split(",")] for l in f]
7    f.close()
8    return lines
9
10
11def euclideanDistance(a, b):
12    return sum([(a[i] - b[i])**2 for i in range(len(a))])**.5
13
14
15def euclideanDistanceWeighted(a, b, w):
16    return sum([(a[i] - b[i])**2 * w[i] for i in range(len(a))])**.5
17
18
19def getNeighborData(dataSet, instance):
20    return sorted([(euclideanDistance(instance, i), i) for i in dataSet])[0][-1]
21
22
23def getNeighborDataWeighted(dataSet, instance, weights):
24    return sorted([(euclideanDistance(instance, i, weights), i) for i in dataSet])[0][-1]
25
26
27def getNeighbors(dataSet, instance, k):
28    return sorted([(euclideanDistance(instance, i), i[-1]) for i in dataSet])[:k]
29
30
31def getNeighborsWeighted(dataSet, instance, k, weights):
32    return sorted([(euclideanDistanceWeighted(instance, i, weights), i[-1]) for i in dataSet])[:k]
33
34
35def predictLabel(dataSet, instance, k):
36    return Counter([i[-1] for i in getNeighbors(dataSet, instance, k)]).most_common()[0][0]
37
38
39def predictLabelWeighted(dataSet, instance, k, weights):
40    return Counter([i[-1] for i in getNeighborsWeighted(dataSet, instance, k, weights)]).most_common()[0][0]