def fit(self,data,eps,MinPts):
X = data
n = X.shape[0]
dist = computeEuclideanDistance(X)#distancematrix
neighborhoods = [np.where(x<=eps)[0] for x in dist]#find all neighbourhoods within eps distance
C = 0
D = data
n = D.shape[0]
visited = [False]* n
# labels = [None]*n
labels = -np.ones(n)#initialise labels to all -1
for index in range(0,n):
if visited[index] == True:
continue
visited[index] = True
NeighbourPts = self.regionQuery(index,eps,neighborhoods)
if len(NeighbourPts) < MinPts:
labels[index] = -1
else:
C += 1
self.expandCluster(index,NeighbourPts,C,eps,MinPts,visited,labels,neighborhoods)
return labels
def plotKnn(self,X,k):
D = computeEuclideanDistance(X);
sorted = np.sort(D)
s = sorted[:,1:k]
kDist = np.sort(np.mean(s[:,1:4],axis = 1))
objectId = np.arange(0,len(kDist))
plt.plot(objectId,kDist)
plt.show()
