Please use this identifier to cite or link to this item:
Title: Air quality health index prediction based on hybrid CNN+LSTM model
Authors: Zhang, Shilin
Keywords: Engineering::Electrical and electronic engineering
Issue Date: 2022
Publisher: Nanyang Technological University
Source: Zhang, S. (2022). Air quality health index prediction based on hybrid CNN+LSTM model. Final Year Project (FYP), Nanyang Technological University, Singapore.
Project: A2254-211
Abstract: With increasing concerns about urban sustainability, air pollution prediction based on environmental monitoring data variables has become more important, providing a reference for industry and people's daily lives. This project aims to develop a supervised model to predict Air Quality Health Index (AQHI) by using actual sensor data and transferring this model between different administrative regions(stations). The model can also be used to predict other air pollutants. This project used a combination of convolution neural network (CNN) and the long short-term memory neural network (LSTM) model to predict the AQHI at multiple locations in the city. The model can predict the data of the next day based on the data of the past seven days or the data of the next hour based on the data of past 24 hours. This project is implemented on the data of Air Quality Health Index (AQHI), Fine Suspended Particulates (FSP), Sulphur Dioxide (SO2), Nitrogen dioxide (NO2), and Respirable Suspended Particulates (RSP) in central and western Hong Kong. The whole model construction process includes adding a CNN layer to the standard LSTM model to extract data features, comparing univariate input and multivariate input, adjusting the data period from daily to hourly, and adjusting the hyperparameters. The source is open data from the Hong Kong Environmental Protection Department (EPD) website. In transfer learning, transfer the network weights of source station to the training model of target station. A Graphical User Interface (GUI) facilitates prediction using the model.
Schools: School of Electrical and Electronic Engineering 
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
3.64 MBAdobe PDFView/Open

Page view(s)

Updated on Sep 21, 2023


Updated on Sep 21, 2023

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.