Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Air Quality Prediction Using the Fractional Gradient-Based Recurrent Neural Network.

Sugandha Arora1, Narinderjit Singh Sawaran Singh2, Divyanshu Singh1

  • 1Birla Institute of Technology and Science Pilani, Pilani, Rajasthan, India.

Computational Intelligence and Neuroscience
|December 19, 2022
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Triple-Band Reconfigurable Monopole Antenna for Long-Range IoT Applications.

Sensors (Basel, Switzerland)·2023
See all related articles

This study enhances air quality index (AQI) prediction in Indian cities using a vanilla recurrent neural network (RNN) with fractional derivatives. The novel approach improves accuracy, showing results comparable to advanced models like LSTM.

Area of Science:

  • Environmental Science
  • Data Science
  • Artificial Intelligence

Background:

  • Air Quality Index (AQI) is crucial for assessing air pollution levels, influenced by factors like weather, traffic, and industrialization.
  • AQI prediction is a time-series problem, historically dependent on past air quality data and environmental conditions.
  • Recurrent Neural Networks (RNNs) are suitable for time-series prediction due to their inherent memory capabilities.

Purpose of the Study:

  • To predict the Air Quality Index (AQI) in Indian cities using a vanilla Recurrent Neural Network (RNN).
  • To investigate the effectiveness of incorporating fractional derivatives into the RNN training process for improved AQI prediction.
  • To compare the performance of the proposed fractional RNN model with traditional RNNs and Long Short-Term Memory (LSTM) networks.

Related Experiment Videos

Main Methods:

  • Utilized a vanilla Recurrent Neural Network (RNN) architecture for AQI prediction.
  • Integrated fractional derivatives, specifically Caputo's derivative, into the gradient descent algorithm for RNN backpropagation.
  • Employed fractional calculus to enhance the memory property of the RNN, capturing historical dependencies in air quality data.

Main Results:

  • The proposed fractional RNN model demonstrated higher accuracy in predicting AQI and key pollutant gases across different Indian cities.
  • The inclusion of fractional derivatives significantly improved the predictive performance of the vanilla RNN.
  • The accuracy achieved by the fractional RNN was found to be comparable to that of more complex models like Long Short-Term Memory (LSTM).

Conclusions:

  • Fractional derivatives can effectively enhance the memory and predictive accuracy of vanilla RNNs for time-series data like AQI.
  • The proposed fractional RNN approach offers a promising and accurate method for air quality forecasting.
  • This study validates the potential of fractional calculus in improving deep learning models for environmental monitoring and prediction.