nokados

import keras.backend as K


def recall(y_true, y_pred):
    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))
    possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))
    recall = true_positives / (possible_positives + K.epsilon())
    return recall

def precision(y_true, y_pred):

nokados

import numpy as np
from gensim.models import KeyedVectors, Word2Vec
from gensim.models.fasttext import FastText as FT_gensim
from nltk.tokenize import sent_tokenize, word_tokenize
import json
import pandas as pd
from tqdm import trange

W2V_PATH = 'data/GoogleNews-vectors-negative300.bin'

nokados

%%time
clusters = dbscan.fit(doc2vec_list)

cl_labels = clusters.labels_.tolist()

def wordcloud_cluster_byIds(cluId):
    texts = []
    for i in range(0, len(cl_labels)):
        if cl_labels[i] == cluId:
            for word in word_tokenize(dialogs_concatted.iloc[i].TEXT):

nokados

def calc_embedding(text):
    tokens = word_tokenize(text)
    vec = np.zeros(100)
    num_tokens = 0
    for token in tokens:
        if token in stopwords_list:
            continue
        if token in new_model:
            vec += new_model[token]
            num_tokens += 1

nokados

import string
translator = str.maketrans('', '', re.sub(r'[\?-]', '', string.punctuation+'«»”“', flags=re.MULTILINE))
def clear_punctuation(sentence):
    return sentence.translate(translator)

nokados

from tqdm import tqdm_notebook
tqdm_notebook().pandas()