

In [1]:

    
%%HTML
<style>
div.cell {
    background-color: rgba(255,200,255,0.85);  
    font-family: Georgia, Cursive;
    font-size: 120%;
    border-radius: 10px;
}

.MathJax_Display {
    text-align: center;
}


body {
    background-color: #AAFFFF;
    background-image: url('https://raw.github.com/chenleo/Resource/master/Images/FirstKiss1600X1200.jpg');
    background-size: cover;
    background-repeat: no-repeat;
    background--attachment: fixed;
}
</style>

Collective Intelligence

Resource:



In [1]:

    
import os
import sys



In [2]:

    
%qtconsole



In [3]:

    
sys.version
#ipython-2.7 notebook (in office)









    Out[3]:





'2.7.6 (default, Nov 12 2013, 13:10:34) \n[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)]'



In [4]:

    
pwd









    Out[4]:





u'/Users/chenchen/Documents/git/ipynotebook'



In [6]:

    
os.chdir("../programming-collective-intelligence-code/")



In [7]:

    
pwd









    Out[7]:





u'/Users/chenchen/Documents/git/programming-collective-intelligence-code'



In [8]:

    
%timeit 1+1









    



10000000 loops, best of 3: 36 ns per loop

Chapter 2: Making Recoomendations



In [9]:

    
os.chdir("chapter2/")



In [10]:

    
!ls









    



deliciousrec.py     pydelicious.py      recommendations.py  recommendations.pyc



In [11]:

    
from recommendations import critics
import recommendations

Finding Similar Users

Euclidean distance
Pearson Collelation

Euclidean Distance

$Dist(P,Q) = {1 \over {1+\sum{(Q_i - P_i)^2}}}$

Dist = 1: same
Dist = 0: different



In [16]:

    
people = critics.keys()
print people









    



['Jack Matthews', 'Mick LaSalle', 'Claudia Puig', 'Lisa Rose', 'Toby', 'Gene Seymour', 'Michael Phillips']



In [17]:

    
recommendations.sim_distance(critics, people[3], people[5])









    Out[17]:





0.14814814814814814

Person Correlation Score



In [ ]: