Sentiment Analysis with BERT Pytorch【半成品】

https://curiousily.com/posts/sentiment-analysis-with-bert-and-hugging-face-using-pytorch-and-python/

导入各种包，设置基本参数

import transformers
from transformers import BertModel, BertTokenizer, AdamW, get_linear_schedule_with_warmup
import torch
import numpy as np
import pandas as pd
import seaborn as sns
from pylab import rcParams
import matplotlib.pyplot as plt
from matplotlib import rc
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix, classification_report
from collections import defaultdict
from textwrap import wrap
from torch import nn, optim
from torch.utils.data import Dataset, DataLoader
%matplotlib inline
%config InlineBackend.figure_format='retina'
sns.set(style='whitegrid', palette='muted', font_scale=1.2)
HAPPY_COLORS_PALETTE = ["#01BEFE", "#FFDD00", "#FF7D00", "#FF006D", "#ADFF02", "#8F00FF"]
sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))
rcParams['figure.figsize'] = 12, 8
RANDOM_SEED = 42
np.random.seed(RANDOM_SEED)
torch.manual_seed(RANDOM_SEED)
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

使用的是Google Play 的评论数据，用下面的程序，从谷歌drive上下载数据集

import gdown
url1 = "https://drive.google.com/uc?id=1S6qMioqPJjyBLpLVz4gmRTnJHnjitnuV"
url2 = 'https://drive.google.com/uc?id=1zdmewp7ayS4js4VtrJEHzAheSW-5NBZv'
gdown.download(url1)  #'apps.csv'
gdown.download(url2)  #'reviews.csv'

读取数据

df = pd.read_csv("reviews.csv")
df.head()

有这些类：

list(df.columns)

['userName',
 'userImage',
 'content',
 'score',
 'thumbsUpCount',
 'reviewCreatedVersion',
 'at',
 'replyContent',
 'repliedAt',
 'sortOrder',
 'appId']

最主要的还是content和score，分别是评论的内容和打分(1-5分)

查看数据形状

df.shape

(15746, 11)

---

查看缺失值情况

df.info()

可以看到content和score没有缺失值，其他的并不影响

---

查看各个打分的分布情况

sns.countplot(df['score'])
plt.xlabel('review score')

相当不平衡，所以考虑将其转化为negative，neutral，positive类，进行情感分析

---

将socre 转化为0，1，2

def to_sentiment(rating):
  rating = int(rating)
  if rating <= 2:
    return 0 # 负面
  elif rating == 3:
    return 1 #中立