Added transformation module containing the standardization function f…

…or feature scaling used by the linreg module. Added algorithms subpackage to mlearn main package.

mlearn/linear_regression.py → mlearn/algorithms/linreg.py

@@ -23,7 +23,7 @@ def sum_of_squared_residuals(x, y, beta):
     ssr : numpy.array
 
     References
-    --------
+    ----------
     .. [1] https://en.wikipedia.org/wiki/Residual_sum_of_squares
 
     Notes

mlearn/algorithms/transform.py

@@ -0,0 +1,60 @@
+"""This module contains general transformation functions."""
+
+import numpy as np
+
+
+def standardize(x,):
+    """
+    Perform z-transformation to get standard score of input matrix.
+
+    Parameters
+    ----------
+    x : numpy.array
+        Input matrix to be standardized.
+
+    Returns
+    -------
+    x_stand : numpy.array
+        Standardized matrix.
+
+    Notes
+    -----
+    .. math:: z = \\frac{x - \\mu}{\\sigma}
+
+    References
+    ----------
+    .. [1] https://en.wikipedia.org/wiki/Standard_score
+
+    Examples
+    --------
+    >>> # The input matrix
+    >>> x = np.array([[1, 11, 104],
+                      [1, 15, 99],
+                      [1, 22, 89],
+                      [1, 27, 88]])
+    >>> standardize(x)
+    [[ 0.         -1.25412576  1.33424877]
+     [ 0.         -0.60683505  0.59299945]
+     [ 0.          0.52592371 -0.88949918]
+     [ 0.          1.3350371  -1.03774904]]
+
+    """
+
+    mean = np.mean(x, axis=0)
+    std = np.std(x, axis=0)
+
+    # handle division by 0
+    with np.errstate(divide='ignore', invalid='ignore'):
+        x_stand = np.true_divide(x-mean,std)
+        x_stand[x_stand == np.inf] = 0
+        x_stand = np.nan_to_num(x_stand)
+
+    return x_stand
+
+if __name__ == "__main__":
+    x = np.array([[1, 11, 104],
+                  [1, 15, 99],
+                  [1, 22, 89],
+                  [1, 27, 88]])
+
+    print(standardize(x))