One Hot Encoding using numpy [duplicate]

by

Usually, when you want to get a one-hot encoding for classification in machine learning, you have an array of indices.

import numpy as np
nb_classes = 6
targets = np.array([[2, 3, 4, 0]]).reshape(-1)
one_hot_targets = np.eye(nb_classes)[targets]

The one_hot_targets is now

array([[[ 0.,  0.,  1.,  0.,  0.,  0.],
        [ 0.,  0.,  0.,  1.,  0.,  0.],
        [ 0.,  0.,  0.,  0.,  1.,  0.],
        [ 1.,  0.,  0.,  0.,  0.,  0.]]])

The .reshape(-1) is there to make sure you have the right labels format (you might also have [[2], [3], [4], [0]]). The -1 is a special value which means “put all remaining stuff in this dimension”. As there is only one, it flattens the array.

Copy-Paste solution

def get_one_hot(targets, nb_classes):
    res = np.eye(nb_classes)[np.array(targets).reshape(-1)]
    return res.reshape(list(targets.shape)+[nb_classes])

Package

You can use mpu.ml.indices2one_hot. It’s tested and simple to use:

import mpu.ml
one_hot = mpu.ml.indices2one_hot([1, 3, 0], nb_classes=5)

More Related Contents:

  1. Convert array of indices to 1-hot encoded numpy array
  2. Convert array of indices to one-hot encoded array in NumPy
  3. Plotting a 2D heatmap with Matplotlib
  4. best way to preserve numpy arrays on disk
  5. How do I catch a numpy warning like it’s an exception (not just for testing)?
  6. How to convert a NumPy array to PIL image applying matplotlib colormap
  7. How to save and load numpy.array() data properly?
  8. Using moviepy, scipy and numpy in amazon lambda
  9. python how to pad numpy array with zeros
  10. Finding the consecutive zeros in a numpy array
  11. Numpy ValueError: setting an array element with a sequence. This message may appear without the existing of a sequence?
  12. Numpy: Fix array with rows of different lengths by filling the empty elements with zeros
  13. Is there a multi-dimensional version of arange/linspace in numpy?
  14. Matplotlib figure to image as a numpy array
  15. Why np.hypot and np.subtract.outer very fast compared to vanilla broadcast ? Using Numba for speedup numpy in parallel for distance matrix calculation
  16. Faster way of polygon intersection with shapely
  17. Find elements of array one nearest to elements of array two
  18. How to convert list of numpy arrays into single numpy array?
  19. how to perform max/mean pooling on a 2d array using numpy
  20. find and delete from more-dimensional numpy array
  21. Initialise numpy array of unknown length
  22. Saving Matplotlib graphs to image as full screen
  23. Create new column based on values from other columns / apply a function of multiple columns, row-wise in Pandas
  24. Filtering (reducing) a NumPy Array
  25. How to replicate array to specific length array
  26. Cannot get minor grid lines to appear in matplotlib figure
  27. Automatic White Balancing with Grayworld assumption
  28. C++ conversion from NumPy array to Mat (OpenCV)
  29. Numpy shuffle multidimensional array by row only, keep column order unchanged
  30. What is the Big O Complexity of Reversing the Order of Columns in Pandas DataFrame?

Leave a Comment