tensorflow for clustering

[lemonysoft]

Is there anyone using tensorflow for clustering? Which dataset is good for a beginner? There is no exercise in ZTM.

[UmbertoFasci]

For a beginner clustering project you can simply create your own data:

import tensorflow as tf
import numpy as np

points_n = 200
clusters_n = 3

points = tf.constant(np.random.uniform(0, 10, (points_n, 2)))
centroids = tf.Variable(tf.slice(tf.random_shuffle(points), [0, 0], [clusters_n, -1]))

This generates random data points and assigns them to a 2D-tensor constant. Also, it randomly chooses centroids from the same generated set.

There is a lot more that can be done with this like enabling element-wise subtraction between points and centroids by expanding their dimensions. 0, and 1 respectively. With this you can calculate distances between points and centroids and determine what makes a cluster.