In [7]:

    

%cd ~/NetBeansProjects/ExpLosion/
from notebooks.common_imports import *


    

    




/Users/miroslavbatchkarov/NetBeansProjects/ExpLosion

In [2]:

    

from discoutils.thesaurus_loader import Vectors
from random import sample


    

In [3]:

    

v = Vectors.from_tsv('/lustre/scratch/inf/mmb28/FeatureExtractionToolkit/word2vec_vectors/word2vec-wiki-nopos-15perc.unigr.strings.rep0')


    

In [5]:

    

v.init_sims(n_neighbors=5)


    

In [ ]:

    

words = random.sample(list(v.keys()), 10)
data = []
for w in words:
    neigh = v.get_nearest_neighbours(w)
    data.append([n[0] for n in neigh[:5]])


    

In [19]:

    

df = pd.DataFrame(data, index=words, columns=['Neigh %d'%d for d in range(1, 6)])


    

In [20]:

    

df


    

    
Out[20]:






  

    

      

      
Neigh 1
      
Neigh 2
      
Neigh 3
      
Neigh 4
      
Neigh 5
    
  
  

    

      
godmother
      
betrothed
      
lover
      
spinster
      
mother-in-law
      
stepdaughter
    
    

      
performer
      
musician
      
artist
      
entertainer
      
dancer
      
accompanist
    
    

      
confiscation
      
expropriation
      
forfeiture
      
taking
      
curtailment
      
illegality
    
    

      
amal
      
al-amin
      
mukhtar
      
nasrallah
      
usama
      
ashraf
    
    

      
alms
      
penitent
      
passer-by
      
sainthood
      
yagna
      
repentant
    
    

      
pawnbroker
      
thrifty
      
bookie
      
yusuke
      
fleer
      
buchalter
    
    

      
plas
      
cupar
      
vanden
      
bodoni
      
eastlake
      
suk
    
    

      
menuhin
      
yehudi
      
caryl
      
barenboim
      
barone
      
lalo
    
    

      
parent
      
mother
      
family
      
live-in
      
child
      
carer
    
    

      
roden
      
coy
      
kempthorne
      
orme
      
porteous
      
allred
    
  

In [21]:

    

print(df.to_latex())


    

    




\begin{tabular}{llllll}
\toprule
{} &        Neigh 1 &     Neigh 2 &      Neigh 3 &        Neigh 4 &       Neigh 5 \\
\midrule
godmother    &      betrothed &       lover &     spinster &  mother-in-law &  stepdaughter \\
performer    &       musician &      artist &  entertainer &         dancer &   accompanist \\
confiscation &  expropriation &  forfeiture &       taking &    curtailment &    illegality \\
amal         &        al-amin &     mukhtar &    nasrallah &          usama &        ashraf \\
alms         &       penitent &   passer-by &    sainthood &          yagna &     repentant \\
pawnbroker   &        thrifty &      bookie &       yusuke &          fleer &     buchalter \\
plas         &          cupar &      vanden &       bodoni &       eastlake &           suk \\
menuhin      &         yehudi &       caryl &    barenboim &         barone &          lalo \\
parent       &         mother &      family &      live-in &          child &         carer \\
roden        &            coy &  kempthorne &         orme &       porteous &        allred \\
\bottomrule
\end{tabular}