Natural Language Processing¶

---

This example shows how to use ATOM to quickly go from raw text data to model predictions.

Import the 20 newsgroups text dataset from sklearn.datasets. The dataset comprises around 18000 articles on 20 topics. The goal is to predict the topic of every article.

Load the data¶

In [1]:

import numpy as np
from atom import ATOMClassifier
from sklearn.datasets import fetch_20newsgroups

import numpy as np from atom import ATOMClassifier from sklearn.datasets import fetch_20newsgroups

In [2]:

# Use only a subset of the available topics for faster processing
X_text, y_text = fetch_20newsgroups(
    return_X_y=True,
    categories=[
        'alt.atheism',
        'sci.med',
        'comp.windows.x',
        'misc.forsale',
        'rec.autos',
    ],
    shuffle=True,
    random_state=1,
)
X_text = np.array(X_text).reshape(-1, 1)

# Use only a subset of the available topics for faster processing X_text, y_text = fetch_20newsgroups( return_X_y=True, categories=[ 'alt.atheism', 'sci.med', 'comp.windows.x', 'misc.forsale', 'rec.autos', ], shuffle=True, random_state=1, ) X_text = np.array(X_text).reshape(-1, 1)

Run the pipeline¶

In [3]:

atom = ATOMClassifier(X_text, y_text, index=True, test_size=0.3, verbose=2, random_state=1)

atom = ATOMClassifier(X_text, y_text, index=True, test_size=0.3, verbose=2, random_state=1)

<< ================== ATOM ================== >>
Algorithm task: multiclass classification.

Dataset stats ==================== >>
Shape: (2846, 2)
Memory: 5.15 MB
Scaled: False
Categorical features: 1 (100.0%)
-------------------------------------
Train set size: 1993
Test set size: 853
-------------------------------------
|   |     dataset |       train |        test |
| - | ----------- | ----------- | ----------- |
| 0 |   480 (1.0) |   336 (1.0) |   144 (1.0) |
| 1 |   593 (1.2) |   415 (1.2) |   178 (1.2) |
| 2 |   585 (1.2) |   410 (1.2) |   175 (1.2) |
| 3 |   594 (1.2) |   416 (1.2) |   178 (1.2) |
| 4 |   594 (1.2) |   416 (1.2) |   178 (1.2) |

In [4]:

atom.dataset  # Note that the feature is automatically named 'corpus'

atom.dataset # Note that the feature is automatically named 'corpus'

Out[4]:

	corpus	target
2283	From: geb@cs.pitt.edu (Gordon Banks)\nSubject:...	4
961	From: keith@cco.caltech.edu (Keith Allan Schne...	0
796	From: joes@telxon.mis.telxon.com (Joe Staudt)\...	3
832	From: rog@cdc.hp.com (Roger Haaheim)\nSubject:...	4
978	From: forman@ide.com (Bonnie Forman)\nSubject:...	2
...	...	...
2021	From: pmoloney@maths.tcd.ie (Paul Moloney)\nSu...	0
2606	From: jmiller@network.com (Jeff J. Miller)\nSu...	3
2320	From: maher@kong.gsfc.nasa.gov (552)\nSubject:...	1
513	From: beck@irzr17.inf.tu-dresden.de (Andre Bec...	1
733	From: yozzo@watson.ibm.com (Ralph Yozzo)\nSubj...	4

2846 rows × 2 columns

In [5]:

# Let's have a look at the first document
atom.corpus[0]

# Let's have a look at the first document atom.corpus[0]

Out[5]:

'From: Donald Mackie <Donald_Mackie@med.umich.edu>\nSubject: Re: Barbecued foods and health risk\nOrganization: UM Anesthesiology\nLines: 13\nDistribution: world\nNNTP-Posting-Host: 141.214.86.38\nX-UserAgent: Nuntius v1.1.1d9\nX-XXDate: Mon, 19 Apr 93 20:12:06 GMT\n\nIn article <1993Apr18.175802.28548@clpd.kodak.com> Rich Young,\nyoung@serum.kodak.com writes:\n\nStuff deleted\n\n>\t ... have to\n>\t consume unrealistically large quantities of barbecued meat at a\n>\t time."\n\nI have to confess that this is one of my few unfulfilled ambitions.\nNo matter how much I eat, it still seems realistic.\n\nDon Mackie - his opinion\n'

In [6]:

# Clean the documents from noise (emails, numbers, etc...)
atom.textclean()

# Clean the documents from noise (emails, numbers, etc...) atom.textclean()

Cleaning the corpus...
 --> Decoding unicode characters to ascii.
 --> Converting text to lower case.
 --> Dropping 10012 emails from 2830 documents.
 --> Dropping 0 URL links from 0 documents.
 --> Dropping 2214 HTML tags from 1304 documents.
 --> Dropping 2 emojis from 1 documents.
 --> Dropping 31222 numbers from 2843 documents.
 --> Dropping punctuation from the text.

In [7]:

# Have a look at the removed items
atom.drops

# Have a look at the removed items atom.drops

Out[7]:

	email	url	html	emoji	number
2283	[geb@cs.pitt.edu, geb@cs.pitt.edu, mo2@access....	NaN	[<1q6rie$>]	NaN	[15, 1, 5, 400]
961	[keith@cco.caltech.edu, livesey@solntze.wpd.sg...	NaN	NaN	NaN	[16]
796	[joes@telxon.mis.telxon.com, 1993apr20.142818....	NaN	[<>]	NaN	[45, 225, 54, 225, 2, 5582, 44334, 0582, 216, ...
832	[rog@cdc.hp.com, ls8139@albnyvms.bitnet]	NaN	NaN	NaN	[15, 1, 1]
978	[forman@ide.com, forman@ide.com]	NaN	NaN	NaN	[13, 2, 4, 40, 1, 800, 00, 510, 947, 6987]
...	...	...	...	...	...
194	NaN	NaN	NaN	NaN	[15, 1, 1, 1097, 08836, 908, 563, 9033, 908, 5...
1035	NaN	NaN	NaN	NaN	[47, 252, 4, 179, 34, 12]
604	NaN	NaN	NaN	NaN	[15, 1, 1]
711	NaN	NaN	NaN	NaN	[13, 93, 212, 274, 0646, 1097, 08836, 908, 563...
1000	NaN	NaN	NaN	NaN	[38, 84]

2846 rows × 5 columns

In [8]:

# Check how the first document changed
atom.corpus[0]

# Check how the first document changed atom.corpus[0]

Out[8]:

'from donald mackie \nsubject re barbecued foods and health risk\norganization um anesthesiology\nlines \ndistribution world\nnntppostinghost \nxuseragent nuntius v11d9\nxxxdate mon  apr   gmt\n\nin article  rich young\n writes\n\nstuff deleted\n\n\t  have to\n\t consume unrealistically large quantities of barbecued meat at a\n\t time\n\ni have to confess that this is one of my few unfulfilled ambitions\nno matter how much i eat it still seems realistic\n\ndon mackie  his opinion\n'

In [9]:

# Convert the strings to a sequence of words
atom.tokenize()

# Convert the strings to a sequence of words atom.tokenize()

Tokenizing the corpus...

In [10]:

# Print the first few words of the first document
atom.corpus[0][:7]

# Print the first few words of the first document atom.corpus[0][:7]

Out[10]:

['from', 'donald', 'mackie', 'subject', 're', 'barbecued', 'foods']

In [11]:

# Normalize the text to a predefined standard
atom.textnormalize(stopwords="english", lemmatize=True)

# Normalize the text to a predefined standard atom.textnormalize(stopwords="english", lemmatize=True)

Normalizing the corpus...
 --> Dropping stopwords.
 --> Applying lemmatization.

In [12]:

atom.corpus[0][:7]  # Check changes...

atom.corpus[0][:7] # Check changes...

Out[12]:

['donald', 'mackie', 'subject', 'barbecue', 'food', 'health', 'risk']

In [13]:

# Visualize the most common words with a wordcloud
atom.plot_wordcloud()

# Visualize the most common words with a wordcloud atom.plot_wordcloud()

In [14]:

# Have a look at the most frequent bigrams
atom.plot_ngrams(2)

# Have a look at the most frequent bigrams atom.plot_ngrams(2)

In [15]:

# Create the bigrams using the tokenizer
atom.tokenize(bigram_freq=215)

# Create the bigrams using the tokenizer atom.tokenize(bigram_freq=215)

Tokenizing the corpus...
 --> Creating 10 bigrams on 4178 locations.

In [16]:

atom.bigrams

atom.bigrams

Out[16]:

	bigram	frequency
0	x_x	1169
1	line_article	714
2	line_nntppostinghost	493
3	organization_university	367
4	gordon_bank	266
5	line_distribution	258
6	distribution_world	249
7	distribution_usa	229
8	usa_line	217
9	computer_science	216

In [17]:

# As a last step before modelling, convert the words to vectors
atom.vectorize(strategy="tfidf")

# As a last step before modelling, convert the words to vectors atom.vectorize(strategy="tfidf")

Fitting Vectorizer...
Vectorizing the corpus...

In [18]:

# The dimensionality of the dataset has increased a lot!
atom.shape

# The dimensionality of the dataset has increased a lot! atom.shape

Out[18]:

(2846, 27589)

In [19]:

# Note that the data is sparse and the columns are named
# after the words they are embedding
atom.dtypes

# Note that the data is sparse and the columns are named # after the words they are embedding atom.dtypes

Out[19]:

corpus_00          Sparse[float64, 0]
corpus_000         Sparse[float64, 0]
corpus_000000e5    Sparse[float64, 0]
corpus_00000ee5    Sparse[float64, 0]
corpus_000010af    Sparse[float64, 0]
                          ...        
corpus_zvi         Sparse[float64, 0]
corpus_zx          Sparse[float64, 0]
corpus_zyklonb     Sparse[float64, 0]
corpus_zzzs        Sparse[float64, 0]
target                          int64
Length: 27589, dtype: object

In [20]:

# When the dataset is sparse, stats() shows the density
atom.stats()

# When the dataset is sparse, stats() shows the density atom.stats()

Dataset stats ==================== >>
Shape: (2846, 27589)
Memory: 3.18 MB
Sparse: True
Density: 0.33%
-------------------------------------
Train set size: 1993
Test set size: 853
-------------------------------------
|   |     dataset |       train |        test |
| - | ----------- | ----------- | ----------- |
| 0 |   480 (1.0) |   336 (1.0) |   144 (1.0) |
| 1 |   593 (1.2) |   415 (1.2) |   178 (1.2) |
| 2 |   585 (1.2) |   410 (1.2) |   175 (1.2) |
| 3 |   594 (1.2) |   416 (1.2) |   178 (1.2) |
| 4 |   594 (1.2) |   416 (1.2) |   178 (1.2) |

In [21]:

# Check which models have support for sparse matrices
atom.available_models()[["acronym", "fullname", "accepts_sparse"]]

# Check which models have support for sparse matrices atom.available_models()[["acronym", "fullname", "accepts_sparse"]]

Out[21]:

	acronym	fullname	accepts_sparse
0	Dummy	Dummy Estimator	False
1	GP	Gaussian Process	False
2	GNB	Gaussian Naive Bayes	False
3	MNB	Multinomial Naive Bayes	True
4	BNB	Bernoulli Naive Bayes	True
5	CatNB	Categorical Naive Bayes	True
6	CNB	Complement Naive Bayes	True
7	Ridge	Ridge Estimator	True
8	Perc	Perceptron	False
9	LR	Logistic Regression	True
10	LDA	Linear Discriminant Analysis	False
11	QDA	Quadratic Discriminant Analysis	False
12	KNN	K-Nearest Neighbors	True
13	RNN	Radius Nearest Neighbors	True
14	Tree	Decision Tree	True
15	Bag	Bagging	True
16	ET	Extra-Trees	True
17	RF	Random Forest	True
18	AdaB	AdaBoost	True
19	GBM	Gradient Boosting Machine	True
20	hGBM	HistGBM	False
21	XGB	XGBoost	True
22	LGB	LightGBM	True
23	CatB	CatBoost	True
24	lSVM	Linear SVM	True
25	kSVM	Kernel SVM	True
26	PA	Passive Aggressive	True
27	SGD	Stochastic Gradient Descent	True
28	MLP	Multi-layer Perceptron	True

In [22]:

# Train the model
atom.run(models="RF", metric="f1_weighted")

# Train the model atom.run(models="RF", metric="f1_weighted")

Training ========================= >>
Models: RF
Metric: f1_weighted


Results for Random Forest:
Fit ---------------------------------------------
Train evaluation --> f1_weighted: 1.0
Test evaluation --> f1_weighted: 0.9295
Time elapsed: 23.832s
-------------------------------------------------
Total time: 23.832s


Final results ==================== >>
Duration: 23.832s
-------------------------------------
Random Forest --> f1_weighted: 0.9295

Analyze results¶

In [23]:

atom.evaluate()

atom.evaluate()

Out[23]:

	balanced_accuracy	f1_weighted	jaccard_weighted	matthews_corrcoef	precision_weighted	recall_weighted
RF	0.931065	0.929523	0.869258	0.912317	0.930918	0.92966

In [24]:

atom.plot_confusion_matrix(figsize=(10, 10))

atom.plot_confusion_matrix(figsize=(10, 10))

In [25]:

atom.decision_plot(index=0, target=atom.predict(0), show=15)

atom.decision_plot(index=0, target=atom.predict(0), show=15)

In [26]:

atom.beeswarm_plot(target=0, show=15)

atom.beeswarm_plot(target=0, show=15)

100%|===================| 4261/4265 [04:22<00:00]