NMF

from sklearn.decomposition import TruncatedSVD
svd = TruncatedSVD(n_components=100, random_state=1234)
svd.fit(dtm)

import matplotlib.pyplot as plt
plt.plot(svd.explained_variance_)

from sklearn.decomposition import NMF
NUM_TOPICS = 14
nmf = NMF(n_components=NUM_TOPICS)
doc_emb = nmf.fit_transform(dtm)

word_emb = nmf.components_.T

words = cv.get_feature_names_out().tolist()
i = words.index('모발')

plt.plot(word_emb[i])

from sklearn.metrics.pairwise import cosine_similarity
sim = cosine_similarity(word_emb)


import numpy as np
s = np.argsort(sim[i])
related = s[-2:-12:-1]
for j in related:
    print(words[j])

indices = []
target = ['모발', '손상', '두피', '모공',
          '용기', '내용물']
for w in target:
    i = words.index(w)
    indices.append(i)
    print(w, i)

dist = 1 - sim[indices, ][:, indices]

from sklearn.manifold import MDS
mds = MDS(dissimilarity='precomputed', random_state=1234)
pos = mds.fit_transform(dist)

!pip install adjusttext

from adjustText import adjust_text
plt.plot(pos[:, 0], pos[:, 1], '.')
texts = [plt.text(pos[i, 0], pos[i, 1], w) for i, w in enumerate(target)]
adjust_text(texts)

for t in range(NUM_TOPICS):
    print(t)
    topic_words_idx = np.argsort(word_emb[:, t])

    for j in topic_words_idx[-1:-11:-1]:
        print(words[j])

doc_id = 0
df.iloc[doc_id]

plt.plot(doc_emb[doc_id])

topic_id = np.argmax(doc_emb[doc_id])
topic_id

topic_docs_idx = np.argsort(doc_emb[:,topic_id])[-1:-11:-1]
df.iloc[topic_docs_idx]

sims = cosine_similarity(doc_emb[[doc_id]], doc_emb).flatten()
sim_idx = np.argsort(sims)[-1:-11:-1]
df.iloc[sim_idx]

sims[sim_idx]