Loading plotting and (hopefully) modelling of whisker movements



In [19]:

    
import numpy as np
# import tables as tb
import pandas as pd
# import scipy.io as sio
# from scipy import linalg
# import itertools
from scipy import signal
from lpproj import LocalityPreservingProjection
from wpca import WPCA, EMPCA, PCA
# from sklearn.decomposition import PCA
# from sklearn.cluster import KMeans
# from sklearn import mixture
# from sklearn.externals.six.moves import xrange

%matplotlib inline
import matplotlib.pyplot as plt
# import sklearn as sk
import seaborn as sns

from lightning import Lightning
# lgn = Lightning(host="http://localhost:3000",ipython=True)
# lgn.create_session("whiskfree1")
lgn = Lightning(local=True,ipython=True)









    




Lightning initialized






    











    



Running local mode, some functionality limited.



In [4]:

    
theta = np.loadtxt('/Users/mathew/GoogleDrive/Data/Gutnisky_final/behaviour/theta_N18.csv')



In [5]:

    
plt.plot(theta)









    Out[5]:





[<matplotlib.lines.Line2D at 0x10fb0ad90>]



In [6]:

    
theta_bits = np.reshape(theta,(len(theta)/100,100))



In [14]:



In [7]:

    
theta_bits.shape









    Out[7]:





(24100, 100)



In [20]:

    
prange = range(17000,17010)
lgn.line(theta_bits[prange])









    Out[20]:



In [47]:

Try high-pass filtering, then thresholding theta before computing autocorrelation



In [51]:

    
# Jake VDP's different PCAs example function
def plot_results(ThisPCA, X, weights=None, Xtrue=None, ncomp=2,prange=range(10)):
    # Compute the standard/weighted PCA
    if weights is None:
        kwds = {}
    else:
        kwds = {'weights': weights}
    
    # Compute the PCA vectors & variance
    pca = ThisPCA(n_components=10).fit(X, **kwds)
    
    # Reconstruct the data using the PCA model
    Y = ThisPCA(n_components=ncomp).fit_reconstruct(X, **kwds)
        
    # Create the plots
    fig, ax = plt.subplots(2, 2, figsize=(16, 6))
    if Xtrue is not None:
        ax[0, 0].plot(Xtrue[prange].T, c='gray', lw=1)
        ax[1, 1].plot(Xtrue[prange].T, c='gray', lw=1)
    ax[0, 0].plot(X[prange].T, c='black', lw=1)
    ax[1, 1].plot(Y[prange].T, c='black', lw=1)
    
    ax[0, 1].plot(pca.components_[:ncomp].T, c='black')
    
    ax[1, 0].plot(np.arange(1, 11), pca.explained_variance_ratio_)
    ax[1, 0].set_xlim(1, 10)
    ax[1, 0].set_ylim(0, None)
    
    ax[0, 0].xaxis.set_major_formatter(plt.NullFormatter())
    ax[0, 1].xaxis.set_major_formatter(plt.NullFormatter())
    
    ax[0, 0].set_title('Input Data')
    ax[0, 1].set_title('First {0} Principal Vectors'.format(ncomp))
    ax[1, 1].set_title('Reconstructed Data ({0} components)'.format(ncomp))
    ax[1, 0].set_title('PCA variance ratio')
    ax[1, 0].set_xlabel('principal vector')
    ax[1, 0].set_ylabel('proportion of total variance')
    
    fig.suptitle(ThisPCA.__name__, fontsize=16)



In [54]:

    
plot_results(PCA, theta_bits,ncomp=5,prange=range(17000,17010))



In [55]:

    
for method in [WPCA, EMPCA]:
    plot_results(method, theta_bits,ncomp=5,prange = range(17000,17010))



In [ ]:

    
# Jake VDP Locality preserving projection example
Xpca = PCA(n_components=3).fit_transform(theta_bits)
Xlpp = LocalityPreservingProjection(n_components=3).fit_transform(theta_bits)



In [58]:

    
fig, ax = plt.subplots(1, 2, figsize=(16, 5))
ax[0].scatter(Xlpp[:, 1], Xlpp[:, 2])
ax[0].set_title('LPP with outliers')
ax[1].scatter(Xpca[:, 1], Xpca[:, 2])
ax[1].set_title('PCA with outliers');



In [16]:

    
#TSNE visualisation
tsne = manifold.TSNE(n_components=2,learning_rate=500,verbose=1,random_state=0)
t0 = time()
X_tsne_pc = tsne.fit_transform(Xpca)
t1 = time()
t1-t0
X_tsne = tsne.fit_transform(theta_bits)
t2 = time()
t2-t1









    



[t-SNE] Computing pairwise distances...
[t-SNE] Computing 91 nearest neighbors...
[t-SNE] Computed conditional probabilities for sample 1000 / 24100
[t-SNE] Computed conditional probabilities for sample 2000 / 24100
[t-SNE] Computed conditional probabilities for sample 3000 / 24100
[t-SNE] Computed conditional probabilities for sample 4000 / 24100
[t-SNE] Computed conditional probabilities for sample 5000 / 24100
[t-SNE] Computed conditional probabilities for sample 6000 / 24100
[t-SNE] Computed conditional probabilities for sample 7000 / 24100
[t-SNE] Computed conditional probabilities for sample 8000 / 24100
[t-SNE] Computed conditional probabilities for sample 9000 / 24100
[t-SNE] Computed conditional probabilities for sample 10000 / 24100
[t-SNE] Computed conditional probabilities for sample 11000 / 24100
[t-SNE] Computed conditional probabilities for sample 12000 / 24100
[t-SNE] Computed conditional probabilities for sample 13000 / 24100
[t-SNE] Computed conditional probabilities for sample 14000 / 24100
[t-SNE] Computed conditional probabilities for sample 15000 / 24100
[t-SNE] Computed conditional probabilities for sample 16000 / 24100
[t-SNE] Computed conditional probabilities for sample 17000 / 24100
[t-SNE] Computed conditional probabilities for sample 18000 / 24100
[t-SNE] Computed conditional probabilities for sample 19000 / 24100
[t-SNE] Computed conditional probabilities for sample 20000 / 24100
[t-SNE] Computed conditional probabilities for sample 21000 / 24100
[t-SNE] Computed conditional probabilities for sample 22000 / 24100
[t-SNE] Computed conditional probabilities for sample 23000 / 24100
[t-SNE] Computed conditional probabilities for sample 24000 / 24100
[t-SNE] Computed conditional probabilities for sample 24100 / 24100
[t-SNE] Mean sigma: 0.000000
[t-SNE] Error after 50 iterations with early exaggeration: 0.870486
[t-SNE] Error after 200 iterations: 0.869738
[t-SNE] Computing pairwise distances...
[t-SNE] Computing 91 nearest neighbors...
[t-SNE] Computed conditional probabilities for sample 1000 / 24100
[t-SNE] Computed conditional probabilities for sample 2000 / 24100
[t-SNE] Computed conditional probabilities for sample 3000 / 24100
[t-SNE] Computed conditional probabilities for sample 4000 / 24100
[t-SNE] Computed conditional probabilities for sample 5000 / 24100
[t-SNE] Computed conditional probabilities for sample 6000 / 24100
[t-SNE] Computed conditional probabilities for sample 7000 / 24100
[t-SNE] Computed conditional probabilities for sample 8000 / 24100
[t-SNE] Computed conditional probabilities for sample 9000 / 24100
[t-SNE] Computed conditional probabilities for sample 10000 / 24100
[t-SNE] Computed conditional probabilities for sample 11000 / 24100
[t-SNE] Computed conditional probabilities for sample 12000 / 24100
[t-SNE] Computed conditional probabilities for sample 13000 / 24100
[t-SNE] Computed conditional probabilities for sample 14000 / 24100
[t-SNE] Computed conditional probabilities for sample 15000 / 24100
[t-SNE] Computed conditional probabilities for sample 16000 / 24100
[t-SNE] Computed conditional probabilities for sample 17000 / 24100
[t-SNE] Computed conditional probabilities for sample 18000 / 24100
[t-SNE] Computed conditional probabilities for sample 19000 / 24100
[t-SNE] Computed conditional probabilities for sample 20000 / 24100
[t-SNE] Computed conditional probabilities for sample 21000 / 24100
[t-SNE] Computed conditional probabilities for sample 22000 / 24100
[t-SNE] Computed conditional probabilities for sample 23000 / 24100
[t-SNE] Computed conditional probabilities for sample 24000 / 24100
[t-SNE] Computed conditional probabilities for sample 24100 / 24100
[t-SNE] Mean sigma: 0.000000
[t-SNE] Error after 50 iterations with early exaggeration: 0.845006
[t-SNE] Error after 225 iterations: 0.844266






    Out[16]:





717.4649369716644



In [14]:

    
from sklearn import manifold
from time import time
t0 = time()
Xpca.shape
# manifold.TSNE?
t1 = time()
t1-t0









    Out[14]:





8.511543273925781e-05



In [9]:

    
Xpca = PCA(n_components=10).fit_transform(theta_bits)



In [18]:

    
fig, ax = plt.subplots(1,2,figsize=(16,6))
ax[0].scatter(X_tsne[:,0],X_tsne[:,1])
ax[1].scatter(X_tsne_pc[:,0],X_tsne_pc[:,1])









    Out[18]:





<matplotlib.collections.PathCollection at 0x111a60290>






    



/Users/mathew/anaconda/lib/python2.7/site-packages/matplotlib/collections.py:590: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
  if self._edgecolors == str('face'):



In [24]:

    
# lgn = Lightning(host="http://localhost:3000",ipython=False)
# lgn.scatter(X_tsne[:,0],X_tsne[:,1],size=1,colormap='Purples')
# lgn.scatter?
mds = manifold.MDS(2, max_iter=100, n_init=1)
X_mds = mds.fit_transform(Xpca)



In [25]:

    
plt.figure(figsize=(16,16))
plt.scatter(X_tsne_pc[:,0],X_tsne_pc[:,1])









    Out[25]:





<matplotlib.collections.PathCollection at 0x11a662750>



In [29]:

    
fig, ax = plt.subplots(1,2,figsize=(16,6))
ax[0].scatter(X_tsne_pc[:,0],X_tsne_pc[:,1])
ax[1].scatter(X_mds[:,0],X_mds[:,1])









    Out[29]:





<matplotlib.collections.PathCollection at 0x11b5a0390>



In [30]:

    
acorr = signal.fftconvolve(theta, theta[::-1], mode='full') # much faster than np.correlate
lgn.line(acorr[len(theta)-1:len(theta)+10000])









    Out[30]:



In [31]:

    
Y = PCA(n_components=10).fit_reconstruct(theta_bits)



In [100]:

    
from pandas import DataFrame
df = DataFrame(Xpca)



In [112]:

    
# Try seaborn pairplot of first 5 PCs
import random
# sns.pairplot?
# x = df.sample(100)
# x?
random.sample(df.index, 10)









    Out[112]:





[19965, 2969, 13231, 20051, 13310, 9887, 10545, 22324, 17945, 23417]



In [124]:

    
x = Xpca[random.sample(df.index, 1000),:5]
# sns.pairplot(df[random.sample(df.index, 10),:5])
df_r = DataFrame(x)
sns.pairplot(df_r)









    Out[124]:





<seaborn.axisgrid.PairGrid at 0x71f720510>



In [89]:

    
lgn.line([np.reshape(Y,(len(theta)))[5e5:6e5],theta[5e5:6e5]])
# lgn.line()









    



/Users/mathew/anaconda/lib/python2.7/site-packages/ipykernel/__main__.py:1: DeprecationWarning: using a non-integer number instead of an integer will result in an error in the future
  if __name__ == '__main__':






    Out[89]:



In [91]:



In [ ]:



In [ ]:



In [ ]: