jcheong0428

# Windowed time lagged cross correlation
seconds = 5
fps = 30
no_splits = 20
samples_per_split = df.shape[0]/no_splits
rss=[]
for t in range(0, no_splits):
    d1 = df['S1_Joy'].loc[(t)*samples_per_split:(t+1)*samples_per_split]
    d2 = df['S2_Joy'].loc[(t)*samples_per_split:(t+1)*samples_per_split]
    rs = [crosscorr(d1,d2, lag) for lag in range(-int(seconds*fps),int(seconds*fps+1))]

jcheong0428

from dtw import dtw,accelerated_dtw

d1 = df['S1_Joy'].interpolate().values
d2 = df['S2_Joy'].interpolate().values
d, cost_matrix, acc_cost_matrix, path = accelerated_dtw(d1,d2, dist='euclidean')

plt.imshow(acc_cost_matrix.T, origin='lower', cmap='gray', interpolation='nearest')
plt.plot(path[0], path[1], 'w')
plt.xlabel('Subject1')
plt.ylabel('Subject2')

jcheong0428

from scipy.signal import hilbert, butter, filtfilt
from scipy.fftpack import fft,fftfreq,rfft,irfft,ifft
import numpy as np
import seaborn as sns
import pandas as pd
import scipy.stats as stats
def butter_bandpass(lowcut, highcut, fs, order=5):
    nyq = 0.5 * fs
    low = lowcut / nyq
    high = highcut / nyq

jcheong0428

jcheong0428

jcheong0428

jcheong0428

# First gist for Hungarian post.
# Brute force estimation of optimal assignment for given preference. 
%matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_context('talk')
sns.set_style('white')
import itertools
import numpy as np

jcheong0428

%%timeit
# Second gist for Hungarian post.
# Estimates the time needed for brute force optimal assignment for group size 10. 
# Takes about 3 minutes.
maxsums, colixs = [], []
group_size = 10
np.random.seed(0)
preferences = np.random.rand(group_size,group_size)

rowix = list(range(group_size)) # rows don't need to be permuted

jcheong0428

%%timeit
# Third gist for Hungarian post.
# Estimates the time needed for Hungarian algorithm optimal assignment for group size 10. 
# Hungarian algorithm below is adopted from Scipy: https://github.com/scipy/scipy/blob/v0.18.1/scipy/optimize/_hungarian.py
###########
# Hungarian algorithm (Kuhn-Munkres) for solving the linear sum assignment
# problem. Taken from scikit-learn. Based on original code by Brian Clapper,
# adapted to NumPy by Gael Varoquaux.
# Further improvements by Ben Root, Vlad Niculae, and Lars Buitinck.
#

jcheong0428

require(MASS)
# Generage fake data for male
data_male <- data.frame(mvrnorm(n=1000,mu=c(2,0),Sigma=rbind(c(1,.8),c(.8,1)),empirical=TRUE ) )
colnames(data_male)<-c('Income','Age')
data_male$Gender = 'Male'
# Generate fake data for female
data_female <- data.frame(mvrnorm(n=1000,mu=c(3,0),Sigma=rbind(c(1,.3),c(.3,1)),empirical=TRUE ))
colnames(data_female) <- c('Income','Age')
data_female$Gender = 'Female'
# Combine data