Merge 2feb072 into 060450a

examples/plot_stacked_interp.py

@@ -0,0 +1,111 @@
+'''
+===========================
+Resampling Stacked Time Series Data
+===========================
+This is a basic example illustrating the resampling of stacked format time series data
+This may be useful for resampling irregularly sampled time series in stacked format, or for determining
+an optimal sampling frequency for the data
+'''
+# Author: Phil Boyer
+# License: BSD
+
+import matplotlib.pyplot as plt
+import numpy as np
+
+from seglearn.datasets import load_stacked_data
+from seglearn.transform import StackedInterp
+
+print(">> Example 1 <<")
+
+# Example 1 - Simple stacked input with values from 2 sensors / 2 axis at irregular sample times
+
+t = np.array([1.1, 1.2, 2.1, 3.3, 3.4, 3.5]).astype(float)
+s = np.array([0, 1, 0, 0, 1, 1]).astype(float)
+v1 = np.array([3, 4, 5, 7, 15, 25]).astype(float)
+v2 = np.array([5, 7, 6, 9, 22, 35]).astype(float)
+y = np.array([1, 2, 2, 2, 3, 3]).astype(float)
+df = np.column_stack([t, s, v1, v2])
+
+X = [df, df]
+y = [y, y]
+
+print("\nX input = " + str(X))
+
+stacked_interp = StackedInterp(0.5)
+stacked_interp.fit(X, y)
+Xc, yc, swt = stacked_interp.transform(X, y)
+
+print ("\nX interpolated: " + str(Xc))
+
+print("\n>> Example 2 <<")
+
+# Example 2 - Stacked input with 3 sensors / 3 axis in 2 time series at irregular sample times
+
+# Boolean: 1 if input data is in nanoseconds - 0 if not
+inNanoseconds = 1
+
+# seed RNGESUS
+np.random.seed(123124)
+
+# load the data
+X = load_stacked_data()
+
+#print("data = " + str(X))
+
+N = len(X)
+
+# I am adding in a column to represent targets (y) since my data doesn't include it
+y = [np.array(np.arange(len(X[i])) + np.random.rand(len(X[i]))).astype(float) for i in np.arange(N)]
+
+# Plot the 2 time series, each of the sensors versus the new sample period for 2 sample periods
+sample_periods = [(10. / 100.)*(inNanoseconds*10**9), (1. / 100.)*(inNanoseconds*10**9), (0.1 / 100.)*(inNanoseconds*10**9)]
+
+f, axarr = plt.subplots(2, 3)
+
+# Matplotlib style parameters
+left = 0.125  # the left side of the subplots of the figure
+right = 0.9  # the right side of the subplots of the figure
+bottom = 0.1  # the bottom of the subplots of the figure
+top = 0.9  # the top of the subplots of the figure
+wspace = 0.3  # the amount of width reserved for space between subplots,
+# expressed as a fraction of the average axis width
+hspace = 0.4  # the amount of height reserved for space between subplots,
+# expressed as a fraction of the average axis height
+
+plt.subplots_adjust(left=left, bottom=bottom, right=right, top=top, wspace=wspace, hspace=hspace)
+
+for j in np.arange(len(sample_periods)):
+    print("\nSample Period = " + str(sample_periods[j] / (inNanoseconds * 10 ** 9)) + " s")
+    stacked_interp = StackedInterp(sample_periods[j])
+    stacked_interp.fit(X, y)
+    Xc, yc, swt = stacked_interp.transform(X, y)
+
+    for i in np.arange(N):
+
+        print("X[" + str(i) + "] length = " + str(len(X[i])))
+        print("X[" + str(i) + "] length after interpolation to sample period = " + str(len(Xc[i])))
+
+        axarr[i, j].plot(Xc[i])
+        axarr[i, j].set_title("InterpSeries " + str(i) + ", P = "
+                              + str(sample_periods[j]/(inNanoseconds*10**9)) + " s")
+        axarr[i, j].set_xlabel("Sample Number")
+        axarr[i, j].set_ylabel("Value")
+
+        # set the grid on
+        axarr[i, j].grid('on')
+
+        # adjust axis label font
+        xlab = axarr[i, j].xaxis.get_label()
+        ylab = axarr[i, j].yaxis.get_label()
+
+        xlab.set_style('italic')
+        xlab.set_size(8)
+        ylab.set_style('italic')
+        ylab.set_size(8)
+
+        # adjust title font
+        ttl = axarr[i, j].title
+        ttl.set_weight('bold')
+        ttl.set_size('8')
+
+plt.show()