Scribd
Upload
16 views4 pages

Text Classification Reseach Paper

xdcfvgbhnjmk

Uploaded by

Manish jaiswal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
16 views4 pages

Text Classification Reseach Paper

xdcfvgbhnjmk

Uploaded by

Manish jaiswal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Proceedings of the 3d International Conference Computational Linguistics And Intelligent Systems

 extending the structure (addition of elements) and vice versa;


 splitting long sentences into several smaller ones and vice versa;
 replacing synonymous words orphrases (collocations) [2].
The paper proposesthe method for paraphrase extraction from the news text
corpus based on the developed syntactic rules [3] to define phrases (collocations) and
the use of WordNet [4] to identify synonymous words in the text corpus.
The developed corpus consists ofBBCnews articles, the sport section [5].
For preprocessing(POS-tagging), the NLTK's Python language library tools are
suggested to use.
Figure 1 shows the synonymous pairs obtained in WordNet.

Fig. 1. Synonymous Pairs Extracted from WordNet

For extracting paraphrases, we check the correspondence of the grammatical


characteristics of collocates (synonymous words of phrases identified at the previous
stage) with the syntactic rules.
Thus, phrases whose grammatical characteristics correspond to the rules are
considered to be synonymous.As a result, the proposed method for paraphrase
extraction from the news text corpus allows identifying a common information space
for topical news.

References
1. Koloiev, A.S.: Rewrite as a new phenomenon in modern journalism. In: SPU Bulletin.
Philology, vol. 1, 221-226 (2012)
2. Bolshakov, I.A.: Two methods of synonymous paraphrasing in linguistic steganography.
In:Proceedings of the International ConferenceDialogue-2004,http://www.dialog-
21.ru/media/2496/bolshakov.pdf, last accessed 2019/02/10.
3. Petrasova, S., Khairova, N., Lewoniewski, W.: Building the semantic similarity model for
social network data streams. In:Data Stream Mining & Processing, Proceedings of the
2018 IEEE Second International Conference (DSMP), 21-24 (2018)
4. WordNet: https://wordnet.princeton.edu, last accessed 2019/02/10.
5. BBC, https://www.bbc.com/news,last accessed 2019/02/10.

70

COLINS’2019, Volume II: Workshop. Kharkiv, Ukraine, April 18-19,


2019, ISSN 2523-4013 http://colins.in.ua, online
Proceedings of the 3d International Conference Computational Linguistics And Intelligent Systems

Machine Learning Text Classification Model with NLP


Approach

Maria Razno[0000-0003-3356-5027]
National Technical University "Kharkiv Polytechnic Institute",
Pushkinska str., 79/2, Kharkiv, Ukraine

[email protected]

Abstract. This article describes the relevance of the word processing task that
is written in human language by the methods of Machine Learning and NLP
approach, that can be used on Python programming language. It also portrays
the concept of Machine Learning, its main varieties and the most popular
Pythonpackages and libraries for working with text data using Machine
Learning methods. The concept of NLP and the most popular python packages
are also presented in the article. The machine learning classification model
algorithm based on the text processing is introduced in the article. It shows how
to use classification machine learning and NLP methods in practice.

Keywords: Machine learning, Python, Pandas, Text classification, NLP,


NLTK, Scikit-learn, Artificial Intelligence, Python Library, Deep Learning
Texts

Over the last few years machine learning and artificial intelligence have become
very hot topics. Nowadays their methods and approaches are a part of a huge amount
of products, moreover it is a necessary thing in most applications and appliances. An
example of using ML (Machine Learning) can be the automatic determination of
important emails and quick responses in Gmail. Nowadays we can confidently say
that and artificial intelligence with machine learning can push a person out of many
technological processes.
Machine learning is the scientific study of algorithms and statistical methods that
computer systems use to effectively perform a specific task without using explicit
instructions, relying on patterns and inference instead. It is seen as a subset of
artificial intelligence. Machine learning algorithms build a mathematical model of
sample data, known as "training data", in order to make predictions or decisions
without being explicitly programmed to perform the task. There are five types of
machine learning algorithms: supervised, semi-supervised, active learning,
reinforcement and unsupervised learning [1].
Natural language processing is a subfield of computer science, information
engineering, and artificial intelligence concerned with the interactions between
computers and human (natural) languages, in particular, how to program computers in
order to process and analyze large amounts of natural language data. Tasks in natural

71

COLINS’2019, Volume II: Workshop. Kharkiv, Ukraine, April 18-19,


2019, ISSN 2523-4013 http://colins.in.ua, online
Proceedings of the 3d International Conference Computational Linguistics And Intelligent Systems

language processing frequently involve speech recognition, natural language


understanding, and natural language generation.
Text classification is one of the most important and typical task in supervised
machine learning. Assigning categories of documents, which can be a web page,
library book, media articles, gallery etc. has many applications like spam filtering,
email routing, sentiment analysis etc. We would like to demonstrate how we can do
text classification using the most common python machine learning and natural
language processing packages like: Pandas, Scikit-learn, Numpy and little bit of
NLTK.
In our study, we are creating the model, that will be able to classify user`s
comment and give it a star rate from 1 to 5. Supervised machine learning requires to
have prepared labeled data, so we use Yelp_academic_dataset_review in json format.
We downloaded the dataset via the link: https://www.kaggle.com/yelp-dataset/yelp-
dataset#yelp_academic _dataset_review.json. We got a lot of necessary tools by using
Pandas library, that helped us to store data in convenient table, the columns of which
are classification parameters and the rows – information for each classified object.
This form of data storage is very effective in our study, especially for further
accessing a particular column of data during the text processing [2].
The next step is to use natural language processing methods to normalize the text
data. During our work we realized, that the package of libraries NLTK(Natural
Language Toolkit) is great for our purpose. Thanks to its methods we removed all the
stop words, that were not necessary for further analysis, from the text data. Also we
needed to use text stemming in order to remove morphological affixes from the text.
All of those step helped the model to make an accurate analysis of the text data and
get the best clear features for the future classification [3].
The next step of our study was building the machine learning model. We used
Scikit-learn due to the fact, that it is a wonderful library with a huge amount of
opportunities. It has various types of analysis, moreover, it is the most convenient
way of forming a model, because it provides a single interface for all conversion steps
and the final result. Instead of using “Bag of words” approach and counting of each
word in our text data, we use the tf-idf method for each pair of words in our reviews.
Tf-idf normalizes the count by dividing the total sum of the meeting of a certain pair
of words into the number of reviews in which these words appear. In such way, we
get the model, that will find the most common words for each star rate, in other words
it will get the appropriate features and select the best ones. As the result of our study,
the model will be able to analyze user`s comment according to found features.
To summarize, in the course of our research, we can say that Python is a
wonderful programming language, which provides a lot of great libraries for creating
powerful machine learning models and proper natural language processing. For the
task of building a machinelearning text classification model with NLP approach, we
have reviewed the most popular machine learning libraries like : Pandas, Scikit-learn,
Numpy, NLTK and built the text classification model with NLP approach. At the end
of our study we get the learning model, that gives user answer with the appropriate
star rate to user, according to his comment, and the list of the most common words for
each star rate.

72

COLINS’2019, Volume II: Workshop. Kharkiv, Ukraine, April 18-19,


2019, ISSN 2523-4013 http://colins.in.ua, online
Proceedings of the 3d International Conference Computational Linguistics And Intelligent Systems

References
1. Langley, P.: Human and machine learning.Machine Learning,1, pp. 243–248 (1986)
2. Masch, C.: Text classification with Convolution Neural Net-works on Yelp, IMDB &
sentence polarity dataset, https://github.com/cmasch/cnn-text-classification,24/02/2019.
3. Moschitti, A., Basili, R.: Complex Linguistic Features for Text Classification: A
Comprehensive Study. In: Lecture Notes in Computer Science vol. 2997, pp. 181-196,
Springer Science + Business Media (2004)

73

COLINS’2019, Volume II: Workshop. Kharkiv, Ukraine, April 18-19,


2019, ISSN 2523-4013 http://colins.in.ua, online

You might also like