Python Machine Learning 2nd Edition by Sebastian Raschka, Packt Publishing Ltd. 2017
Code Repository: https://github.com/rasbt/python-machine-learning-book-2nd-edition
Code License: MIT License
Note that the optional watermark extension is a small IPython notebook plugin that I developed to make the code reproducible. You can just skip the following line(s).
In [1]:
%load_ext watermark
%watermark -a "Sebastian Raschka" -u -d -v -p numpy,pandas,sklearn,nltk
The use of watermark is optional. You can install this IPython extension via "pip install watermark". For more information, please see: https://github.com/rasbt/watermark.
The IMDB movie review set can be downloaded from http://ai.stanford.edu/~amaas/data/sentiment/. After downloading the dataset, decompress the files.
A) If you are working with Linux or MacOS X, open a new terminal windowm cd into the download directory and execute
tar -zxf aclImdb_v1.tar.gz
B) If you are working with Windows, download an archiver such as 7Zip to extract the files from the download archive.
Optional code to download and unzip the dataset via Python:
In [2]:
import os
import sys
import tarfile
import time
source = 'http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz'
target = 'aclImdb_v1.tar.gz'
def reporthook(count, block_size, total_size):
global start_time
if count == 0:
start_time = time.time()
return
duration = time.time() - start_time
progress_size = int(count * block_size)
speed = progress_size / (1024.**2 * duration)
percent = count * block_size * 100. / total_size
sys.stdout.write("\r%d%% | %d MB | %.2f MB/s | %d sec elapsed" %
(percent, progress_size / (1024.**2), speed, duration))
sys.stdout.flush()
if not os.path.isdir('aclImdb') and not os.path.isfile('aclImdb_v1.tar.gz'):
if (sys.version_info < (3, 0)):
import urllib
urllib.urlretrieve(source, target, reporthook)
else:
import urllib.request
urllib.request.urlretrieve(source, target, reporthook)
In [3]:
if not os.path.isdir('aclImdb'):
with tarfile.open(target, 'r:gz') as tar:
tar.extractall()
In [4]:
import pyprind
import pandas as pd
import os
# change the `basepath` to the directory of the
# unzipped movie dataset
basepath = 'aclImdb'
labels = {'pos': 1, 'neg': 0}
pbar = pyprind.ProgBar(50000)
df = pd.DataFrame()
for s in ('test', 'train'):
for l in ('pos', 'neg'):
path = os.path.join(basepath, s, l)
for file in sorted(os.listdir(path)):
with open(os.path.join(path, file),
'r', encoding='utf-8') as infile:
txt = infile.read()
df = df.append([[txt, labels[l]]],
ignore_index=True)
pbar.update()
df.columns = ['review', 'sentiment']
Shuffling the DataFrame:
In [5]:
import numpy as np
np.random.seed(0)
df = df.reindex(np.random.permutation(df.index))
Optional: Saving the assembled data as CSV file:
In [6]:
df.to_csv('movie_data.csv', index=False, encoding='utf-8')
In [7]:
import pandas as pd
df = pd.read_csv('movie_data.csv', encoding='utf-8')
df.head(3)
Out[7]:
In [8]:
df.shape
Out[8]:
If you have problems with creating the movie_data.csv, you can find a download a zip archive at
https://github.com/rasbt/python-machine-learning-book-2nd-edition/tree/master/code/ch08/
...
By calling the fit_transform method on CountVectorizer, we just constructed the vocabulary of the bag-of-words model and transformed the following three sentences into sparse feature vectors:
In [6]:
import numpy as np
from sklearn.feature_extraction.text import CountVectorizer
count = CountVectorizer()
docs = np.array([
'The sun is shining',
'The weather is sweet',
'The sun is shining, the weather is sweet, and one and one is two'])
bag = count.fit_transform(docs)
Now let us print the contents of the vocabulary to get a better understanding of the underlying concepts:
In [7]:
print(count.vocabulary_)
As we can see from executing the preceding command, the vocabulary is stored in a Python dictionary, which maps the unique words that are mapped to integer indices. Next let us print the feature vectors that we just created:
Each index position in the feature vectors shown here corresponds to the integer values that are stored as dictionary items in the CountVectorizer vocabulary. For example, the rst feature at index position 0 resembles the count of the word and, which only occurs in the last document, and the word is at index position 1 (the 2nd feature in the document vectors) occurs in all three sentences. Those values in the feature vectors are also called the raw term frequencies: tf (t,d)—the number of times a term t occurs in a document d.
In [8]:
print(bag.toarray())
In [9]:
np.set_printoptions(precision=2)
When we are analyzing text data, we often encounter words that occur across multiple documents from both classes. Those frequently occurring words typically don't contain useful or discriminatory information. In this subsection, we will learn about a useful technique called term frequency-inverse document frequency (tf-idf) that can be used to downweight those frequently occurring words in the feature vectors. The tf-idf can be de ned as the product of the term frequency and the inverse document frequency:
$$\text{tf-idf}(t,d)=\text{tf (t,d)}\times \text{idf}(t,d)$$Here the tf(t, d) is the term frequency that we introduced in the previous section, and the inverse document frequency idf(t, d) can be calculated as:
$$\text{idf}(t,d) = \text{log}\frac{n_d}{1+\text{df}(d, t)},$$where $n_d$ is the total number of documents, and df(d, t) is the number of documents d that contain the term t. Note that adding the constant 1 to the denominator is optional and serves the purpose of assigning a non-zero value to terms that occur in all training samples; the log is used to ensure that low document frequencies are not given too much weight.
Scikit-learn implements yet another transformer, the TfidfTransformer, that takes the raw term frequencies from CountVectorizer as input and transforms them into tf-idfs:
In [10]:
from sklearn.feature_extraction.text import TfidfTransformer
tfidf = TfidfTransformer(use_idf=True,
norm='l2',
smooth_idf=True)
print(tfidf.fit_transform(count.fit_transform(docs))
.toarray())
As we saw in the previous subsection, the word is had the largest term frequency in the 3rd document, being the most frequently occurring word. However, after transforming the same feature vector into tf-idfs, we see that the word is is now associated with a relatively small tf-idf (0.45) in document 3 since it is also contained in documents 1 and 2 and thus is unlikely to contain any useful, discriminatory information.
However, if we'd manually calculated the tf-idfs of the individual terms in our feature vectors, we'd have noticed that the TfidfTransformer calculates the tf-idfs slightly differently compared to the standard textbook equations that we de ned earlier. The equations for the idf and tf-idf that were implemented in scikit-learn are:
The tf-idf equation that was implemented in scikit-learn is as follows:
$$\text{tf-idf}(t,d) = \text{tf}(t,d) \times (\text{idf}(t,d)+1)$$While it is also more typical to normalize the raw term frequencies before calculating the tf-idfs, the TfidfTransformer normalizes the tf-idfs directly.
By default (norm='l2'), scikit-learn's TfidfTransformer applies the L2-normalization, which returns a vector of length 1 by dividing an un-normalized feature vector v by its L2-norm:
To make sure that we understand how TfidfTransformer works, let us walk through an example and calculate the tf-idf of the word is in the 3rd document.
The word is has a term frequency of 3 (tf = 3) in document 3, and the document frequency of this term is 3 since the term is occurs in all three documents (df = 3). Thus, we can calculate the idf as follows:
$$\text{idf}("is", d3) = log \frac{1+3}{1+3} = 0$$Now in order to calculate the tf-idf, we simply need to add 1 to the inverse document frequency and multiply it by the term frequency:
$$\text{tf-idf}("is",d3)= 3 \times (0+1) = 3$$
In [11]:
tf_is = 3
n_docs = 3
idf_is = np.log((n_docs+1) / (3+1))
tfidf_is = tf_is * (idf_is + 1)
print('tf-idf of term "is" = %.2f' % tfidf_is)
If we repeated these calculations for all terms in the 3rd document, we'd obtain the following tf-idf vectors: [3.39, 3.0, 3.39, 1.29, 1.29, 1.29, 2.0 , 1.69, 1.29]. However, we notice that the values in this feature vector are different from the values that we obtained from the TfidfTransformer that we used previously. The nal step that we are missing in this tf-idf calculation is the L2-normalization, which can be applied as follows:
As we can see, the results match the results returned by scikit-learn's TfidfTransformer (below). Since we now understand how tf-idfs are calculated, let us proceed to the next sections and apply those concepts to the movie review dataset.
In [12]:
tfidf = TfidfTransformer(use_idf=True, norm=None, smooth_idf=True)
raw_tfidf = tfidf.fit_transform(count.fit_transform(docs)).toarray()[-1]
raw_tfidf
Out[12]:
In [13]:
l2_tfidf = raw_tfidf / np.sqrt(np.sum(raw_tfidf**2))
l2_tfidf
Out[13]:
In [14]:
df.loc[0, 'review'][-50:]
Out[14]:
In [15]:
import re
def preprocessor(text):
text = re.sub('<[^>]*>', '', text)
emoticons = re.findall('(?::|;|=)(?:-)?(?:\)|\(|D|P)',
text)
text = (re.sub('[\W]+', ' ', text.lower()) +
' '.join(emoticons).replace('-', ''))
return text
In [16]:
preprocessor(df.loc[0, 'review'][-50:])
Out[16]:
In [17]:
preprocessor("</a>This :) is :( a test :-)!")
Out[17]:
In [18]:
df['review'] = df['review'].apply(preprocessor)
In [9]:
from nltk.stem.porter import PorterStemmer
porter = PorterStemmer()
def tokenizer(text):
return text.split()
def tokenizer_porter(text):
return [porter.stem(word) for word in text.split()]
In [10]:
tokenizer('runners like running and thus they run')
Out[10]:
In [11]:
tokenizer_porter('runners like running and thus they run')
Out[11]:
In [12]:
import nltk
nltk.download('stopwords')
Out[12]:
In [13]:
from nltk.corpus import stopwords
stop = stopwords.words('english')
[w for w in tokenizer_porter('a runner likes running and runs a lot')[-10:]
if w not in stop]
Out[13]:
Strip HTML and punctuation to speed up the GridSearch later:
In [24]:
X_train = df.loc[:25000, 'review'].values
y_train = df.loc[:25000, 'sentiment'].values
X_test = df.loc[25000:, 'review'].values
y_test = df.loc[25000:, 'sentiment'].values
In [25]:
from sklearn.pipeline import Pipeline
from sklearn.linear_model import LogisticRegression
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import GridSearchCV
tfidf = TfidfVectorizer(strip_accents=None,
lowercase=False,
preprocessor=None)
param_grid = [{'vect__ngram_range': [(1, 1)],
'vect__stop_words': [stop, None],
'vect__tokenizer': [tokenizer, tokenizer_porter],
'clf__penalty': ['l1', 'l2'],
'clf__C': [1.0, 10.0, 100.0]},
{'vect__ngram_range': [(1, 1)],
'vect__stop_words': [stop, None],
'vect__tokenizer': [tokenizer, tokenizer_porter],
'vect__use_idf':[False],
'vect__norm':[None],
'clf__penalty': ['l1', 'l2'],
'clf__C': [1.0, 10.0, 100.0]},
]
lr_tfidf = Pipeline([('vect', tfidf),
('clf', LogisticRegression(random_state=0))])
gs_lr_tfidf = GridSearchCV(lr_tfidf, param_grid,
scoring='accuracy',
cv=5,
verbose=1,
n_jobs=-1)
Important Note about n_jobs
Please note that it is highly recommended to use n_jobs=-1 (instead of n_jobs=1) in the previous code example to utilize all available cores on your machine and speed up the grid search. However, some Windows users reported issues when running the previous code with the n_jobs=-1 setting related to pickling the tokenizer and tokenizer_porter functions for multiprocessing on Windows. Another workaround would be to replace those two functions, [tokenizer, tokenizer_porter], with [str.split]. However, note that the replacement by the simple str.split would not support stemming.
Important Note about the running time
Executing the following code cell may take up to 30-60 min depending on your machine, since based on the parameter grid we defined, there are 22235 + 22235 = 240 models to fit.
If you do not wish to wait so long, you could reduce the size of the dataset by decreasing the number of training samples, for example, as follows:
X_train = df.loc[:2500, 'review'].values
y_train = df.loc[:2500, 'sentiment'].values
However, note that decreasing the training set size to such a small number will likely result in poorly performing models. Alternatively, you can delete parameters from the grid above to reduce the number of models to fit -- for example, by using the following:
param_grid = [{'vect__ngram_range': [(1, 1)],
'vect__stop_words': [stop, None],
'vect__tokenizer': [tokenizer],
'clf__penalty': ['l1', 'l2'],
'clf__C': [1.0, 10.0]},
]
In [ ]:
## @Readers: PLEASE IGNORE THIS CELL
##
## This cell is meant to generate more
## "logging" output when this notebook is run
## on the Travis Continuous Integration
## platform to test the code as well as
## speeding up the run using a smaller
## dataset for debugging
if 'TRAVIS' in os.environ:
gs_lr_tfidf.verbose=2
X_train = df.loc[:250, 'review'].values
y_train = df.loc[:250, 'sentiment'].values
X_test = df.loc[25000:25250, 'review'].values
y_test = df.loc[25000:25250, 'sentiment'].values
In [26]:
gs_lr_tfidf.fit(X_train, y_train)
Out[26]:
In [27]:
print('Best parameter set: %s ' % gs_lr_tfidf.best_params_)
print('CV Accuracy: %.3f' % gs_lr_tfidf.best_score_)
In [28]:
clf = gs_lr_tfidf.best_estimator_
print('Test Accuracy: %.3f' % clf.score(X_test, y_test))
Please note that gs_lr_tfidf.best_score_ is the average k-fold cross-validation score. I.e., if we have a GridSearchCV object with 5-fold cross-validation (like the one above), the best_score_ attribute returns the average score over the 5-folds of the best model. To illustrate this with an example:
In [29]:
from sklearn.linear_model import LogisticRegression
import numpy as np
from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import cross_val_score
np.random.seed(0)
np.set_printoptions(precision=6)
y = [np.random.randint(3) for i in range(25)]
X = (y + np.random.randn(25)).reshape(-1, 1)
cv5_idx = list(StratifiedKFold(n_splits=5, shuffle=False, random_state=0).split(X, y))
cross_val_score(LogisticRegression(random_state=123), X, y, cv=cv5_idx)
Out[29]:
By executing the code above, we created a simple data set of random integers that shall represent our class labels. Next, we fed the indices of 5 cross-validation folds (cv3_idx) to the cross_val_score scorer, which returned 5 accuracy scores -- these are the 5 accuracy values for the 5 test folds.
Next, let us use the GridSearchCV object and feed it the same 5 cross-validation sets (via the pre-generated cv3_idx indices):
In [30]:
from sklearn.model_selection import GridSearchCV
gs = GridSearchCV(LogisticRegression(), {}, cv=cv5_idx, verbose=3).fit(X, y)
As we can see, the scores for the 5 folds are exactly the same as the ones from cross_val_score earlier.
Now, the bestscore attribute of the GridSearchCV object, which becomes available after fitting, returns the average accuracy score of the best model:
In [31]:
gs.best_score_
Out[31]:
As we can see, the result above is consistent with the average score computed the cross_val_score.
In [32]:
cross_val_score(LogisticRegression(), X, y, cv=cv5_idx).mean()
Out[32]:
In [1]:
# This cell is not contained in the book but
# added for convenience so that the notebook
# can be executed starting here, without
# executing prior code in this notebook
import os
import gzip
if not os.path.isfile('movie_data.csv'):
if not os.path.isfile('movie_data.csv.gz'):
print('Please place a copy of the movie_data.csv.gz'
'in this directory. You can obtain it by'
'a) executing the code in the beginning of this'
'notebook or b) by downloading it from GitHub:'
'https://github.com/rasbt/python-machine-learning-'
'book-2nd-edition/blob/master/code/ch08/movie_data.csv.gz')
else:
with in_f = gzip.open('movie_data.csv.gz', 'rb'), \
out_f = open('movie_data.csv', 'wb'):
out_f.write(in_f.read())
In [2]:
import numpy as np
import re
from nltk.corpus import stopwords
# The `stop` is defined as earlier in this chapter
# Added it here for convenience, so that this section
# can be run as standalone without executing prior code
# in the directory
stop = stopwords.words('english')
def tokenizer(text):
text = re.sub('<[^>]*>', '', text)
emoticons = re.findall('(?::|;|=)(?:-)?(?:\)|\(|D|P)', text.lower())
text = re.sub('[\W]+', ' ', text.lower()) +\
' '.join(emoticons).replace('-', '')
tokenized = [w for w in text.split() if w not in stop]
return tokenized
def stream_docs(path):
with open(path, 'r', encoding='utf-8') as csv:
next(csv) # skip header
for line in csv:
text, label = line[:-3], int(line[-2])
yield text, label
In [3]:
next(stream_docs(path='movie_data.csv'))
Out[3]:
In [4]:
def get_minibatch(doc_stream, size):
docs, y = [], []
try:
for _ in range(size):
text, label = next(doc_stream)
docs.append(text)
y.append(label)
except StopIteration:
return None, None
return docs, y
In [5]:
from sklearn.feature_extraction.text import HashingVectorizer
from sklearn.linear_model import SGDClassifier
vect = HashingVectorizer(decode_error='ignore',
n_features=2**21,
preprocessor=None,
tokenizer=tokenizer)
Note
Perceptron(n_iter, ...) by Perceptron(max_iter, ...) in scikit-learn >= 0.19.
In [6]:
from distutils.version import LooseVersion as Version
from sklearn import __version__ as sklearn_version
if Version(sklearn_version) < '0.18':
clf = SGDClassifier(loss='log', random_state=1, n_iter=1)
else:
clf = SGDClassifier(loss='log', random_state=1, max_iter=1)
doc_stream = stream_docs(path='movie_data.csv')
In [7]:
import pyprind
pbar = pyprind.ProgBar(45)
classes = np.array([0, 1])
for _ in range(45):
X_train, y_train = get_minibatch(doc_stream, size=1000)
if not X_train:
break
X_train = vect.transform(X_train)
clf.partial_fit(X_train, y_train, classes=classes)
pbar.update()
In [8]:
X_test, y_test = get_minibatch(doc_stream, size=5000)
X_test = vect.transform(X_test)
print('Accuracy: %.3f' % clf.score(X_test, y_test))
In [9]:
clf = clf.partial_fit(X_test, y_test)
In [1]:
import pandas as pd
df = pd.read_csv('movie_data.csv', encoding='utf-8')
df.head(3)
Out[1]:
In [ ]:
## @Readers: PLEASE IGNORE THIS CELL
##
## This cell is meant to create a smaller dataset if
## the notebook is run on the Travis Continuous Integration
## platform to test the code on a smaller dataset
## to prevent timeout errors and just serves a debugging tool
## for this notebook
if 'TRAVIS' in os.environ:
df.loc[:500].to_csv('movie_data.csv')
df = pd.read_csv('movie_data.csv', nrows=500)
print('SMALL DATA SUBSET CREATED FOR TESTING')
In [2]:
from sklearn.feature_extraction.text import CountVectorizer
count = CountVectorizer(stop_words='english',
max_df=.1,
max_features=5000)
X = count.fit_transform(df['review'].values)
In [3]:
from sklearn.decomposition import LatentDirichletAllocation
lda = LatentDirichletAllocation(n_topics=10,
random_state=123,
learning_method='batch')
X_topics = lda.fit_transform(X)
In [4]:
lda.components_.shape
Out[4]:
In [5]:
n_top_words = 5
feature_names = count.get_feature_names()
for topic_idx, topic in enumerate(lda.components_):
print("Topic %d:" % (topic_idx + 1))
print(" ".join([feature_names[i]
for i in topic.argsort()\
[:-n_top_words - 1:-1]]))
Based on reading the 5 most important words for each topic, we may guess that the LDA identified the following topics:
To confirm that the categories make sense based on the reviews, let's plot 5 movies from the horror movie category (category 6 at index position 5):
In [6]:
horror = X_topics[:, 5].argsort()[::-1]
for iter_idx, movie_idx in enumerate(horror[:3]):
print('\nHorror movie #%d:' % (iter_idx + 1))
print(df['review'][movie_idx][:300], '...')
Using the preceeding code example, we printed the first 300 characters from the top 3 horror movies and indeed, we can see that the reviews -- even though we don't know which exact movie they belong to -- sound like reviews of horror movies, indeed. (However, one might argue that movie #2 could also belong to topic category 1.)
...
Readers may ignore the next cell.
In [10]:
! python ../.convert_notebook_to_script.py --input ch08.ipynb --output ch08.py