NLP: Sentiment Analysis with LSTM

Introduction to Sentiment Analysis

Sentiment Analysis, a common application in Natural Language Processing (NLP), is an interesting task aimed at extracting emotional content from text and categorizing it. It involves analyzing, processing, summarizing, and inferring subjective text with emotional tones.

This article will focus on sentiment polarity analysis within sentiment analysis. Sentiment polarity analysis refers to categorizing text into positive, negative, or neutral sentiments. In most applications, sentiments are classified into two categories. For example, words like “like” and “dislike” represent different emotional orientations.

This article will delve into using the LSTM model, a deep learning model, for sentiment analysis on Chinese text.

Text Introduction and Corpus Analysis

We’ll use comments about a product from an e-commerce website as our corpus (corpus.csv), which can be downloaded from this link. The dataset consists of 4310 comment entries, categorized as “positive” and “negative”. Here are a few entries from the dataset:

1
2
3
4
5
6
7
8
9
10
evaluation,label
用了一段时间，感觉还不错，可以,正面
电视非常好，已经是家里的第二台了。第一天下单，第二天就到本地了，可是物流的人说车坏了，一直催，客服也帮着催，到第三天下午5点才送过来。父母年纪大了，买个大电视画面清晰，趁着耳朵还好使，享受几年。,正面
电视比想象中的大好多，画面也很清晰，系统很智能，更多功能还在摸索中,正面
不错,正面
用了这么多天了，感觉还不错。夏普的牌子还是比较可靠。希望以后比较耐用，现在是考量质量的时候。,正面
物流速度很快，非常棒，今天就看了电视，非常清晰，非常流畅，一次非常完美的购物体验,正面
非常好，客服还特意打电话做回访,正面
物流小哥不错，辛苦了，东西还没用,正面
送货速度快，质量有保障，活动价格挺好的。希望用的久，不出问题。,正面

Following this, we perform a simple analysis on the corpus:

  • Distribution of sentiments in the dataset.
  • Distribution of comment sentence lengths in the dataset.

We use the following Python script to analyze the sentiment distribution and the length distribution of comment sentences.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib import font_manager
from itertools import accumulate

# Set the font for matplotlib plots
my_font = font_manager.FontProperties(fname="/Library/Fonts/Songti.ttc")

# Calculate sentence lengths and their frequency
df = pd.read_csv('./corpus.csv')
print(df.groupby('label')['label'].count())

df['length'] = df['evaluation'].apply(lambda x: len(x))
len_df = df.groupby('length').count()
sent_length = len_df.index.tolist()
sent_freq = len_df['evaluation'].tolist()

# Plot the histogram of sentence lengths and their frequency
plt.bar(sent_length, sent_freq)
plt.title("Histogram of Sentence Lengths and Frequency", fontproperties=my_font)
plt.xlabel("Sentence Length", fontproperties=my_font)
plt.ylabel("Frequency of Sentence Length", fontproperties=my_font)
plt.savefig("./Histogram_of_Sentence_Lengths.png")
plt.close()

# Plot the Cumulative Distribution Function (CDF) of sentence lengths
sent_percentage_list = [(count/sum(sent_freq)) for count in accumulate(sent_freq)]
plt.plot(sent_length, sent_percentage_list)

# Find the sentence length at quantile percentile
quantile = 0.91
for length, per in zip(sent_length, sent_percentage_list):
    if round(per, 2) == quantile:
        index = length
        break
print("\nSentence length at %s quantile: %d." % (quantile, index))

# Plot the CDF of sentence lengths
plt.plot(sent_length, sent_percentage_list)
plt.hlines(quantile, 0, index, colors="c", linestyles="dashed")
plt.vlines(index, 0, quantile, colors="c", linestyles="dashed")
plt.text(0, quantile, str(quantile))
plt.text(index, 0, str(index))
plt.title("Cumulative Distribution Function of Sentence Lengths", fontproperties=my_font)
plt.xlabel("Sentence Length", fontproperties=my_font)
plt.ylabel("Cumulative Frequency of Sentence Length", fontproperties=my_font)
plt.savefig("./CDF_of_Sentence_Lengths.png")
plt.close()

The output result would be:

1
2
3
4
5
6
label
Positive    1908
Negative    2375
Name: label, dtype: int64

Sentence length at quantile 0.91: 183.

Using LSTM Model

Next, we employ the LSTM (Long Short-Term Memory) model from deep learning for sentiment analysis on the provided dataset. The complete Python code for this process is provided:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
# -*- coding: utf-8 -*-

import pickle
import numpy as np
import pandas as pd
from keras.utils import np_utils, plot_model
from keras.models import Sequential
from keras.preprocessing.sequence import pad_sequences
from keras.layers import LSTM, Dense, Embedding, Dropout
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

# 导入数据
# 文件的数据中，特征为evaluation, 类别为label.
def load_data(filepath, input_shape=20):
    df = pd.read_csv(filepath)

    # 标签及词汇表
    labels, vocabulary = list(df['label'].unique()), list(df['evaluation'].unique())

    # 构造字符级别的特征
    string = ''
    for word in vocabulary:
        string += word

    vocabulary = set(string)

    # 字典列表
    word_dictionary = {word: i+1 for i, word in enumerate(vocabulary)}
    with open('word_dict.pk', 'wb') as f:
        pickle.dump(word_dictionary, f)
    inverse_word_dictionary = {i+1: word for i, word in enumerate(vocabulary)}
    label_dictionary = {label: i for i, label in enumerate(labels)}
    with open('label_dict.pk', 'wb') as f:
        pickle.dump(label_dictionary, f)
    output_dictionary = {i: labels for i, labels in enumerate(labels)}

    vocab_size = len(word_dictionary.keys()) # 词汇表大小
    label_size = len(label_dictionary.keys()) # 标签类别数量

    # 序列填充，按input_shape填充，长度不足的按0补充
    x = [[word_dictionary[word] for word in sent] for sent in df['evaluation']]
    x = pad_sequences(maxlen=input_shape, sequences=x, padding='post', value=0)
    y = [[label_dictionary[sent]] for sent in df['label']]
    y = [np_utils.to_categorical(label, num_classes=label_size) for label in y]
    y = np.array([list(_[0]) for _ in y])

    return x, y, output_dictionary, vocab_size, label_size, inverse_word_dictionary

# 创建深度学习模型， Embedding + LSTM + Softmax.
def create_LSTM(n_units, input_shape, output_dim, filepath):
    x, y, output_dictionary, vocab_size, label_size, inverse_word_dictionary = load_data(filepath)
    model = Sequential()
    model.add(Embedding(input_dim=vocab_size + 1, output_dim=output_dim,
                        input_length=input_shape, mask_zero=True))
    model.add(LSTM(n_units, input_shape=(x.shape[0], x.shape[1])))
    model.add(Dropout(0.2))
    model.add(Dense(label_size, activation='softmax'))
    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

    plot_model(model, to_file='./model_lstm.png', show_shapes=True)
    model.summary()

    return model

# 模型训练
def model_train(input_shape, filepath, model_save_path):

    # 将数据集分为训练集和测试集，占比为9:1
    # input_shape = 100
    x, y, output_dictionary, vocab_size, label_size, inverse_word_dictionary = load_data(filepath, input_shape)
    train_x, test_x, train_y, test_y = train_test_split(x, y, test_size = 0.1, random_state = 42)

    # 模型输入参数，需要自己根据需要调整
    n_units = 100
    batch_size = 32
    epochs = 5
    output_dim = 20

    # 模型训练
    lstm_model = create_LSTM(n_units, input_shape, output_dim, filepath)
    lstm_model.fit(train_x, train_y, epochs=epochs, batch_size=batch_size, verbose=1)

    # 模型保存
    lstm_model.save(model_save_path)

    N = test_x.shape[0]  # 测试的条数
    predict = []
    label = []
    for start, end in zip(range(0, N, 1), range(1, N+1, 1)):
        sentence = [inverse_word_dictionary[i] for i in test_x[start] if i != 0]
        y_predict = lstm_model.predict(test_x[start:end])
        label_predict = output_dictionary[np.argmax(y_predict[0])]
        label_true = output_dictionary[np.argmax(test_y[start:end])]
        print(''.join(sentence), label_true, label_predict) # 输出预测结果
        predict.append(label_predict)
        label.append(label_true)

    acc = accuracy_score(predict, label) # 预测准确率
    print('模型在测试集上的准确率为: %s.' % acc)

if __name__ == '__main__':
    filepath = './corpus.csv'
    input_shape = 180
    model_save_path = './corpus_model.h5'
    model_train(input_shape, filepath, model_save_path)

For the aforementioned model, it was trained 5 times with a training-to-testing set ratio of 9:1, yielding the following output:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
......
Epoch 5/5
......
3424/3854 [=========================>....] - ETA: 2s - loss: 0.1280 - acc: 0.9565
3456/3854 [=========================>....] - ETA: 1s - loss: 0.1274 - acc: 0.9569
3488/3854 [==========================>...] - ETA: 1s - loss: 0.1274 - acc: 0.9570
3520/3854 [==========================>...] - ETA: 1s - loss: 0.1287 - acc: 0.9568
3552/3854 [==========================>...] - ETA: 1s - loss: 0.1290 - acc: 0.9564
3584/3854 [==========================>...] - ETA: 1s - loss: 0.1284 - acc: 0.9568
3616/3854 [===========================>..] - ETA: 1s - loss: 0.1284 - acc: 0.9569
3648/3854 [===========================>..] - ETA: 0s - loss: 0.1278 - acc: 0.9572
3680/3854 [===========================>..] - ETA: 0s - loss: 0.1271 - acc: 0.9576
3712/3854 [===========================>..] - ETA: 0s - loss: 0.1268 - acc: 0.9580
3744/3854 [============================>.] - ETA: 0s - loss: 0.1279 - acc: 0.9575
3776/3854 [============================>.] - ETA: 0s - loss: 0.1272 - acc: 0.9579
3808/3854 [============================>.] - ETA: 0s - loss: 0.1279 - acc: 0.9580
3840/3854 [============================>.] - ETA: 0s - loss: 0.1281 - acc: 0.9581
3854/3854 [==============================] - 18s 5ms/step - loss: 0.1298 - acc: 0.9577
......
给父母买的，特意用了一段时间再来评价，电视非常好，没有坏点和损坏，界面也很简洁，便于操作，稍微不足就是开机会比普通电视慢一些，这应该是智能电视的通病吧，如果可以希望微鲸大大可以更新系统优化下开机时间~电视真的很棒，性价比爆棚，值得大家考虑购买。 客服很细心，快递小哥很耐心的等我通电验货，态度非常好。 负面 正面
长须鲸和海狮回答都很及时，虽然物流不够快但是服务不错电视不错，对比了乐视小米和微鲸论性价比还是微鲸好点 负面 负面
所以看不到4k效果，但是应该可以。 自带音响，中规中矩吧，好像没有别人说的好。而且，到现在没连接上我的漫步者，这个非常不满意，因为看到网上说好像普通3.5mm的连不上或者连上了声音小。希望厂家接下来开发的电视有改进。不知道我要不要换个音响。其他的用用再说。 放在地上的是跟我混了两年的tcl，天气受潮，修了一次，下岗了。 最后，我也觉得底座不算太稳，凑合着用。 负面 负面
电视机一般，低端机不要求那么高咯。 负面 负面
很好，两点下单上午就到了，服务很好。 正面 正面
帮朋友买的，好好好好好好好好 正面 正面
......
模型在测试集上的准确率为: 0.9020979020979021.

The model achieved an accuracy of over 95% on the training set and over 90% on the testing set, indicating a fairly good performance.

Model Prediction

Subsequently, we use the trained model to predict the sentiment polarity of new data. The Python code for sentiment prediction is as follows:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
# -*- coding: utf-8 -*-

# Import the necessary modules
import pickle
import numpy as np
from keras.models import load_model
from keras.preprocessing.sequence import pad_sequences


# 导入字典
with open('word_dict.pk', 'rb') as f:
    word_dictionary = pickle.load(f)
with open('label_dict.pk', 'rb') as f:
    output_dictionary = pickle.load(f)

try:
    # 数据预处理
    input_shape = 180
    sent = "电视刚安装好，说实话，画质不怎么样，很差！"
    x = [[word_dictionary[word] for word in sent]]
    x = pad_sequences(maxlen=input_shape, sequences=x, padding='post', value=0)

    # 载入模型
    model_save_path = './sentiment_analysis.h5'
    lstm_model = load_model(model_save_path)

    # 模型预测
    y_predict = lstm_model.predict(x)
    label_dict = {v:k for k,v in output_dictionary.items()}
    print('输入语句: %s' % sent)
    print('情感预测结果: %s' % label_dict[np.argmax(y_predict)])

except KeyError as err:
    print("您输入的句子有汉字不在词汇表中，请重新输入！")
    print("不在词汇表中的单词为：%s." % err)

The output result would be:

1
2
输入语句: 电视刚安装好，说实话，画质不怎么样，很差！
情感预测结果: 负面

Let’s try testing a few other comments:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
输入语句: 物超所值，真心不错
情感预测结果: 正面
输入语句: 很大很好，方便安装！
情感预测结果: 正面
输入语句: 卡，慢，死机，闪退。
情感预测结果: 负面
输入语句: 这种货色就这样吧，别期待怎样。
情感预测结果: 负面
输入语句: 啥服务态度码，出了事情一个推一个，送货安装还收我50
情感预测结果: 负面
输入语句: 京东服务很好！但我买的这款电视两天后就出现这样的问题，很后悔买了这样的电视
情感预测结果: 负面
输入语句: 产品质量不错，就是这位客服的态度十分恶劣，对相关服务不予解释说明，缺乏耐心，
情感预测结果: 负面
输入语句: 很满意，电视非常好。护眼模式，很好，也很清晰。
情感预测结果: 负面

Thank you for reading!

Natural language processing
#NLP
NLP: Sentiment Analysis with LSTM
http://example.com/2023/10/28/NLP-Sentiment-Analysis-with-LSTM/
Author
Jenny Qu
Posted on
October 28, 2023
Licensed under