38 KiB
38 KiB
import pandas as pd
from many_stop_words import get_stop_words
from sklearn.feature_extraction.text import TfidfVectorizer
from unidecode import unidecode
from nltk.tokenize import word_tokenize
import string
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
data=pd.read_csv('dev-0/in.tsv', sep='\t', header=None)
data_test=pd.read_csv('test-A/in.tsv', sep='\t', header=None)
def remove_punctuations(text):
for punctuation in string.punctuation:
text = text.replace(punctuation, '')
return text
data[0] = data[0].str.lower()
data_test[0] = data_test[0].str.lower()
stop_words = get_stop_words('pl')
data[0] = data[0].apply(unidecode)
data_test[0] = data_test[0].apply(unidecode)
uni_stop_words = [unidecode(x) for x in stop_words]
data[0] = data[0].apply(remove_punctuations)
data_test[0] = data_test[0].apply(remove_punctuations)
data[0] = data[0].apply(lambda x: ' '.join([item for item in x.split() if item not in uni_stop_words]))
data_test[0] = data_test[0].apply(lambda x: ' '.join([item for item in x.split() if item not in uni_stop_words]))
tf=TfidfVectorizer()
text_tf= tf.fit_transform(data[0])
text_test_tf= tf.fit_transform(data_test[0])
text_tf
text_test_tf
<691x15352 sparse matrix of type '<class 'numpy.float64'>' with 42571 stored elements in Compressed Sparse Row format>
Sum_of_squared_distances = []
K = range(2,20)
for k in K:
km = KMeans(n_clusters=k, max_iter=200, n_init=10)
km = km.fit(text_tf)
Sum_of_squared_distances.append(km.inertia_)
plt.plot(K, Sum_of_squared_distances, 'bx-')
plt.xlabel('k')
plt.ylabel('Sum_of_squared_distances')
plt.title('Elbow Method For Optimal k')
plt.show()
Sum_of_squared_distances = []
K = range(2,30)
for k in K:
km = KMeans(n_clusters=k, max_iter=200, n_init=10)
km = km.fit(text_test_tf)
Sum_of_squared_distances.append(km.inertia_)
plt.plot(K, Sum_of_squared_distances, 'bx-')
plt.xlabel('k')
plt.ylabel('Sum_of_squared_distances')
plt.title('Elbow Method For Optimal k')
plt.show()
true_k_dev = 10
model_dev = KMeans(n_clusters=true_k_dev, init='k-means++', max_iter=200, n_init=10)
model_dev.fit(text_tf)
labels_dev=model_dev.labels_
clusters_dev=pd.DataFrame(list(labels_dev),columns=['cluster'])
true_k_test = 28
model_test = KMeans(n_clusters=true_k_test, init='k-means++', max_iter=200, n_init=10)
model_test.fit(text_test_tf)
labels_test=model_test.labels_
clusters_test=pd.DataFrame(list(labels_test),columns=['cluster'])
clusters_dev.to_csv("dev-0\out.tsv", sep="\t",index=False,header=None)
clusters_test.to_csv("test-A\out.tsv", sep="\t",index=False,header=None)