Technology NewsTechnology NewsTechnology News
  • Computing
  • AI
  • Robotics
  • Cybersecurity
  • Electric Vehicle
  • Wearables
  • Gaming
  • Space
Reading: Innovative Approach to Arabic Text Classification Using Analogical Proportions
Share
Font ResizerAa
Technology NewsTechnology News
Font ResizerAa
Search
  • Computing
  • AI
  • Robotics
  • Cybersecurity
  • Electric Vehicle
  • Wearables
  • Gaming
  • Space
Follow US
  • Cookie Policy (EU)
  • Contact
  • About
© 2025 NEWSLINKER - Powered by LK SOFTWARE
AIScience News

Innovative Approach to Arabic Text Classification Using Analogical Proportions

Highlights

  • Analogical proportions improve Arabic text classification accuracy.

  • AATC1 and AATC2 outperform traditional methods on several benchmarks.

  • Balanced performance across small and large datasets is achieved.

Samantha Reed
Last updated: 18 June, 2024 - 5:55 am 5:55 am
Samantha Reed 12 months ago
Share
SHARE

A recent article in Expert Systems, EarlyView introduces a novel method for Arabic text classification, leveraging analogical proportions to enhance the accuracy of existing classifiers. The study, titled “Arabic text classification based on analogical proportions,” proposes two new classifiers, AATC1 and AATC2, which outperform traditional techniques such as k-NN, SVM, and even some deep learning approaches, particularly in handling both small and large datasets. This approach offers a promising alternative to current methodologies, aiming to bridge the gap in effectiveness for various dataset sizes.

Contents
Analogical Proportions in Text ClassificationEvaluation and Results

Analogical Proportions in Text Classification

Text classification, a process involving the labeling of text documents with predefined categories, traditionally employs machine learning (ML) algorithms or deep learning techniques. Classic ML algorithms like k-NN and SVM often fall short in accuracy, especially when dealing with small text datasets. Conversely, deep learning models excel with large datasets but struggle with smaller ones involving numerous categories. The current research addresses these limitations by applying analogical proportions to both structured and unstructured text data, effectively bridging the gap between different dataset sizes.

The analogical model proposed in this study expresses relationships between text documents and their real categories through statements such as “x is to y as z is to t.” By developing this model, researchers created two analogical Arabic text classifiers, AATC1 and AATC2. These classifiers predict the category of a new document based on the categories of three other documents from the training set, provided that these four documents form a valid analogical proportion. This innovative method aims to improve classification rates across different text collections.

Evaluation and Results

Extensive experiments were conducted using five benchmark Arabic text collections, including ANT (versions 2.1 and 1.1), BBC-Arabic, CNN-Arabic, and AlKhaleej-2004. The results showed that AATC2 achieved the highest average accuracy of 78.78% and the best average precision of 0.77 on the ANT corpus v2.1. AATC1 also demonstrated superior performance, with the best average precisions of 0.88 and 0.92 for the BBC-Arabic corpus and AlKhaleej-2004, respectively, and an 85.64% average accuracy for CNN-Arabic. These findings highlight the potential of analogical proportions in enhancing text classification accuracy.

Comparing with historical data, previous methods primarily relied on ML algorithms and deep learning techniques. These traditional methods struggled with the dichotomy of handling small versus large datasets effectively. Earlier research illustrated the challenges in achieving high accuracy and precision when classifying Arabic text, particularly with smaller datasets. The introduction of analogical proportions represents a significant shift in methodology, offering a more balanced approach across different dataset sizes.

The current study builds upon these earlier efforts by presenting a methodology that not only improves accuracy but also generalizes well across various dataset sizes. This advancement marks a notable departure from the previous reliance on either classic ML or deep learning techniques alone. The combined use of analogical proportions represents a hybrid approach that integrates the strengths of both traditional and modern classification methods.

The proposed analogical classifiers, AATC1 and AATC2, offer a significant improvement in text classification for Arabic documents. By leveraging the unique properties of analogical proportions, these classifiers outperform several existing methods. This research also underscores the importance of exploring new avenues to enhance text classification accuracy, demonstrating the efficacy of analogical proportions in achieving this goal. The ongoing evolution in text classification methodologies highlights the need for continuous innovation to address the ever-growing and diverse nature of text data.

You can follow us on Youtube, Telegram, Facebook, Linkedin, Twitter ( X ), Mastodon and Bluesky

You Might Also Like

Persona AI Develops Industrial Humanoids to Boost Heavy Industry Work

DeepSeek Restricts Free Speech with R1 0528 AI Model

Grammarly Pursues Rapid A.I. Growth After $1 Billion Funding Boost

AMR Experts Weigh Growth, AI Impact, and Technical Hurdles

Odyssey AI Model Turns Video Into Real-Time Interactive Worlds

Share This Article
Facebook Twitter Copy Link Print
Samantha Reed
By Samantha Reed
Samantha Reed is a 40-year-old, New York-based technology and popular science editor with a degree in journalism. After beginning her career at various media outlets, her passion and area of expertise led her to a significant position at Newslinker. Specializing in tracking the latest developments in the world of technology and science, Samantha excels at presenting complex subjects in a clear and understandable manner to her readers. Through her work at Newslinker, she enlightens a knowledge-thirsty audience, highlighting the role of technology and science in our lives.
Previous Article iPhone Repeatedly Requests Outlook Password: Solutions Explained
Next Article Get Better at Wordle with These Tips

Stay Connected

6.2kLike
8kFollow
2.3kSubscribe
1.7kFollow

Latest News

Tesla Opts for Imports as It Enters Indian Market
Electric Vehicle
Kineis Launches IoT Satellite Services and Enters Asian Markets
IoT
Cadillac Targets Younger Drivers With the New 2025 Optiq Electric SUV
Electric Vehicle
Nvidia Eyes Entry Into Handheld Gaming PC Market With New SoC
Computing
Apple Launches Dedicated Gaming App as WWDC 2025 Approaches
Gaming
NEWSLINKER – your premier source for the latest updates in ai, robotics, electric vehicle, gaming, and technology. We are dedicated to bringing you the most accurate, timely, and engaging content from across these dynamic industries. Join us on our journey of discovery and stay informed in this ever-evolving digital age.

ARTIFICAL INTELLIGENCE

  • Can Artificial Intelligence Achieve Consciousness?
  • What is Artificial Intelligence (AI)?
  • How does Artificial Intelligence Work?
  • Will AI Take Over the World?
  • What Is OpenAI?
  • What is Artifical General Intelligence?

ELECTRIC VEHICLE

  • What is Electric Vehicle in Simple Words?
  • How do Electric Cars Work?
  • What is the Advantage and Disadvantage of Electric Cars?
  • Is Electric Car the Future?

RESEARCH

  • Robotics Market Research & Report
  • Everything you need to know about IoT
  • What Is Wearable Technology?
  • What is FANUC Robotics?
  • What is Anthropic AI?
Technology NewsTechnology News
Follow US
About Us   -  Cookie Policy   -   Contact

© 2025 NEWSLINKER. Powered by LK SOFTWARE
Welcome Back!

Sign in to your account

Register Lost your password?