Machine Learning for Smart Cities: A Survey

2021 IEEE International Power and Renewable Energy Conference (IPRECON), Journal Year: 2022, Volume and Issue: unknown, P. 1 - 8

Published: Dec. 16, 2022

Smart Cities utilize Information and Communication Technology (ICT) tools to improve operational efficiency provide excellent service. It aims make the core infrastructure available enhance quality of life. Artificial Intelligence (AI) approaches are used critical features a smart city cities' sustainable development is needed ensure that rapid urbanization does not affect natural environment. Machine Learning (ML) an essential subset can contribute expansion emerging cities with sustainability. The literature shows research community use Deep (DL) various attributes. These include prediction air quality, crop management, forecasting weather conditions like rainfall, humidity, fog, transportation, water supply, infrastructure, etc. This paper presents literature-based study concept, sustainability in cities, functional aspects survey related it.

Language: Английский

---

Temporal trends and predictive modeling of air pollutants in Delhi: a comparative study of artificial intelligence models

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Dec. 28, 2024

Language: Английский

---

A hybrid model for predicting air quality combining Holt–Winters and Deep Learning Approaches: A novel method to identify ozone concentration peaks

Mathematical Modeling and Computing, Journal Year: 2023, Volume and Issue: 10(4), P. 1154 - 1163

Published: Jan. 1, 2023

Ozone (O3) from the troposphere is one of substances that has a strong effect on air pollution in city Tanger. Prediction this pollutant can have positive improvements quality. This paper presents new approach combining deep-learning algorithms and Holt–Winters method order to detect peaks obtain more accurate forecasting model. Given LSTM an extremely powerful algorithm, we hybridized with enhance Making use multiple accuracy metrics, models' efficiency investigated. Empirical findings reveal superiority hybrid model by providing forecasts are index agreement equal 0.91.

Language: Английский

---

Multistep Ahead Forecasting of Electrical Conductivity in Rivers by Using a Hybrid Convolutional Neural Network-Long Short Term Memory (CNN-LSTM) Model Enhanced by Boruta-XGBoost Feature Selection Algorithm

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: March 16, 2023

Electrical conductivity (EC) is a key water quality metric for predicting the salinity and mineralization. In this study, 10-day-ahead EC of two Australian rivers, Albert River Barratta Creek, was forecasted using novel deep learning algorithm, i.e., convolutional neural network combined with long short-term memory (CNN-LSTM) model. The Boruta-extreme gradient boosting (XGBoost, XGB) feature selection method used to determine significant inputs (time series lagged data) performance proposed Boruta-XGB-CNN-LSTM model compared those three machine approaches: multi-layer perceptron (MLP), K-nearest neighbor (KNN), XGBoost, considering different statistical metrics such as correlation coefficient (R), root mean square error (RMSE), absolute percentage (MAPE). Ten years data both rivers were extracted, seven (2012–2018) (2019–2021) training testing models, respectively. algorithm outperformed other models in forecasting 1-day-ahead stations over test dataset (R = 0.9429, RMSE 45.6896, MAPE 5.9749 River; R 0.9215, 43.8315, 7.6029 Creek). addition, could effectively forecast next 3–10 days. Nevertheless, slightly deteriorated horizon increased from 3 10 Overall, an effective soft computing accurately fluctuation rivers.

Language: Английский

---

A Ensemble Model for PM2.5 Concentration Prediction Based on Feature Selection and Two-Layer Clustering Algorithm

Published: Aug. 18, 2023

Determining accurate PM2.5 pollution concentrations and understanding their dynamic patterns is crucial for scientifically informed air control strategies. Traditional reliance on linear correlation coefficients ascertaining related factors only uncovers superficial relationships. Moreover, the invariance of conventional prediction models restricts accuracy. To enhance precision concentration prediction, this study introduces a novel integrated model that leverages feature selection clustering algorithm. Comprising three components - selection, clustering, first employs non-dominated sorting Genetic Algorithm (NSGA-III) to identify most impactful features affecting within pollutants meteorological factors. This step offers more valuable data subsequent modules. The then adopts two-layer method (SOM+K-means) analyze multifaceted irregularity dataset. Finally, establishes Extreme Learning Machine (ELM) weak learner each classification, integrating multiple learners using Adaboost algorithm obtain comprehensive model. Through enhancement, exploration, adaptability improvement, proposed significantly enhances overall performance. Data sourced from 12 Beijing-based monitoring sites in 2016 were utilized an empirical study, model's results compared with five other predictive models. outcomes demonstrate heightens accuracy, offering useful insights potential broadened application multifactor methodologies pollutants.

Language: Английский

---