def nnmf_analysis():
if len(sys.argv) == 1:
target_year = 1997
else:
target_year = int(sys.argv[1])
fname = str(target_year) + '-' + str(target_year+1) + '.xls'
X, flabels, PLY, ptr1, ptr2, ptr3 = parse_stats(fname)
# pos = {C, F, G}
ptr1 = np.array(X[:,0]==1).transpose()
ptr2 = np.array(X[:,0]==2).transpose()
ptr3 = np.array(X[:,0]==3).transpose()
X = normalize(X)
# NNMF
w, h = nnmf.factorize(X[:,1:], pc=3, iter=100)
print np.shape(h)
print np.shape(w)
print
print h[:,0]
print
print h[:,1]
print
print h[:,2]
print
print h[:,3]
return
def recommend(article, n, nofeatures, max_iter=50):
#Factorize the matrix, make features
word_dictionary, article_words, paper_titles = makeFeatures()
wordmatrix,wordvec = makematrix(word_dictionary, article_words)
plt.plot(sorted(word_dictionary.values()))
plt.savefig("WordFrequencies.png")
w,h = nnmf.factorize(np.matrix(wordmatrix), nofeatures, max_iter)
print "Recommendations for article: "
print article
print "-----------------------------"
tofind = unicode(paper_titles.index(article))
nearest = nn.findnearest(w[tofind,:], w, paper_titles, 1, 0)
nn.printnearest(nearest, 5)
def visualizeFactorization(nofeatures=5, max_iter=10): word_dictionary, article_words, paper_titles = makeFeatures() wordmatrix,wordvec = makematrix(word_dictionary, article_words) w,h = nnmf.factorize(np.matrix(wordmatrix), nofeatures, max_iter) toppatterns=[[] for i in range(len(paper_titles))] patternnames=[] pc, wc = np.shape(h) for i in range(pc): slist=[] # Create a list of words and their weights for j in range(wc): slist.append((h[i,j],wordvec[j])) # Reverse sort the word list slist.sort() slist.reverse() # Print the first six elements n=[s[1] for s in slist[0:10]] print "----------------------------------------" print "Features/Topic Found" print str(i) + str(n) print "Top 3 Articles Closest Matching Feature/Topic" flist=[] for j in range(len(paper_titles)): # Add the article with its weight flist.append((w[j,i],paper_titles[j])) toppatterns[j].append((w[j,i],i,paper_titles[j])) # Reverse sort the list flist.sort() flist.reverse() # Show the top 3 art for f in flist[0:5]: print f
print rows
print
# Extract the volume field from every line
prices[t]=[float(r.split(',')[5]) for r in rows[1:] if r.strip()!='']
if len(prices[t])<shortest: shortest=len(prices[t])
if not dates:
dates=[r.split(',')[0] for r in rows[1:] if r.strip()!='']
l1=[[prices[tickers[i]][j]
for i in range(len(tickers))]
for j in range(shortest)]
w,h=nnmf.factorize(matrix(l1),pc=5)
print "Printing h"
print
print h
print "Printing w"
print w
print "Exiting"
exit()
# Loop over all the features
for i in range(shape(h)[0]):
print "Feature %d" %i
def showarticles(titles, toppatterns, patternnames, out='articles.txt'):
outfile = file(out, 'w')
# Loop over all the articles
for j in range(len(titles)):
outfile.write(titles[j].encode('utf8') + '\n')
# Get the top features for this article and
# reverse sort them
toppatterns[j].sort()
toppatterns[j].reverse()
# Print the top three patterns
for i in range(3):
outfile.write(
str(toppatterns[j][i][0]) + ' ' +
str(patternnames[toppatterns[j][i][1]]) + '\n')
outfile.write('\n')
outfile.close()
#allwords,articlewords,articletitles = getarticlewords()
#makematrix(allw=allwords,articlew=articletitles)
m1 = matrix([[1, 2, 3], [4, 5, 6]])
m2 = matrix([[1, 2], [3, 4], [5, 6]])
print m1 * m2
w, h = nnmf.factorize(m1 * m2, pc=3, iter=100)
