AUTOMATED NEWS CLASSIFICATION USING MACHINE LEARNING

ASOGAN, ARVIN KUMAR (2020) AUTOMATED NEWS CLASSIFICATION USING MACHINE LEARNING. [Final Year Project] (Submitted)

[thumbnail of 17411_Arun Kumar.pdf] PDF
17411_Arun Kumar.pdf
Restricted to Registered users only

Download (1MB)

Abstract

The way of news consumption has changed drastically since the past as 87% of the
respondent of a survey has cited online as their primary news source which makes the
news article more accessible as ever for the readers, but the accessibility may
overwhelm the readers due to the vast amount of news that is published online every
day. Thus, having news classification is important but with the amount of the news
published, we need the help of the machine learning to classify the news article in
which this project was set to do. The objective of this project is to find the best machine
learning model that can be used for the news classification, developing process for
online news extraction and finally developing an automated system to extract and
classify the news articles. Support Vector Machine (SVM) with TF-IDF Vectorization
method was found to be the best machine learning model for news article classification
and online news article extraction process was done using Python script. After that,
extracted articles are used to develop the machine learning model with an accuracy of
89.92%. The developed model was then used for the classification process in the
automated news classification program which was build in Python as well. At the end,
this project has helped to develop an automated news classification system will be
helpful for the readers as they are able to view the news articles that are interesting to
them.

Item Type: Final Year Project
Subjects: Q Science > Q Science (General)
Departments / MOR / COE: Sciences and Information Technology > Computer and Information Sciences
Depositing User: Mr Ahmad Suhairi Mohamed Lazim
Date Deposited: 23 Sep 2021 23:46
Last Modified: 23 Sep 2021 23:46
URI: http://utpedia.utp.edu.my/id/eprint/21687

Actions (login required)

View Item
View Item