Deep Learning Methods Research Articles

Short-Term Prediction of Origin–Destination Passenger Flow in Urban Rail Transit Systems with Multi-Source Data: A Deep Learning Method Fusing High-Dimensional Features

Short-term origin–destination (OD) passenger flow forecasting is crucial for urban rail transit enterprises aiming to optimise transportation products and increase operating income. As there are large-scale OD pairs in an urban rail transit system, OD passenger flow cannot be obtained in real time (temporal hysteresis). Additionally, the distribution characteristics are also complex. Previous studies mainly focus on passenger flow prediction at metro stations, while few methods solve the OD passenger flow prediction problems of an urban rail transit system. In view of this, we propose a novel deep learning method fusing high-dimensional features (HDF-DL) with multi-source data. The HDF-DL method is combined with three modules. The temporal module incorporates the time-varying, trend, and cyclic characteristics of OD passenger flow, while the latest OD passenger flow time sequence (within 1 h) is excluded from the time-varying characteristics. In the spatial module, the K-means and K-shape algorithms are used to classify OD pairs from multiple perspectives and capture the spatial features, reducing the difficulty of OD passenger flow predictions with large-scale and complex characteristics. Weather factors are considered in the external feature module. The HDF-DL method is tested on a large-scale metro system in China, in which eight baseline models are designed. The results show that the HDF-DL method achieves high prediction accuracy across multiple time granularities, with a mean absolute percentage error of about 10%. OD passenger flow in every departure time interval can be predicted with high and stable accuracy, effectively capturing temporal characteristics. The modular design of HDF-DL, which fuses high-dimensional features and employs appropriate neural networks for different data types, significantly reduces prediction errors and outperforms baseline models.

Graph based recurrent network for context specific synthetic lethality prediction.

The concept of synthetic lethality (SL) has been successfully used for targeted therapies. To further explore SL for cancer therapy, identifying more SL interactions with therapeutic potential are essential. Recently, graph neural network-based deep learning methods have been proposed for SL prediction, which reduce the SL search space of wet-lab based methods. However, these methods ignore that most SL interactions depend strongly on genetic context, which limits the application of the predicted results. In this study, we proposed a graph recurrent network-based model for specific context-dependent SL prediction (SLGRN). In particular, we introduced a Graph Recurrent Network-based encoder to acquire a context-specific, low-dimensional feature representation for each node, facilitating the prediction of novel SL. SLGRN leveraged gate recurrent unit (GRU) and it incorporated a context-dependent-level state to effectively integrate information from all nodes. As a result, SLGRN outperforms the state-of-the-arts models for SL prediction. We subsequently validate novel SL interactions under different contexts based on combination therapy or patient survival analysis. Through in vitro experiments and retrospective clinical analysis, we emphasize the potential clinical significance of this context-specific SL prediction model.

ADVANCEMENTS IN ALZHEIMER’S DIAGNOSIS THROUGH MRI USING BAYESIAN CONVOLUTIONAL NEURAL NETWORKS AND VARIATIONAL INFERENCE

Alzheimer’s disease is one of the brain disorders that can be deadly in older. The disease is less treated and less recognized, but Alzheimer’s disease is now a significant public health problem. Early detection of the disease can significantly reduce symptoms. However, the lack of medical personnel makes handling this disease complex. Therefore, an automatic diagnosis of Alzheimer’s disease is needed with a Magnetic Resonance Imaging (MRI) examination to get an accurate diagnosis of the disease. This study classified the type of Alzheimer’s disease with deep learning methods using the Bayesian Convolutional Neural Network (BCNN) and the Variational Inference (VI) technique. It aims to determine image classification and accuracy level at the level of Alzheimer’s disease by using 2,400 brain MRI images, divided into three classes (non-demented, very mild demented, and mild demented) based on severity. The data was acquired from the kaggle.com website. We use a dataset scenario of 80% for training and 20% for testing, 100x100 pixels, kernel size 3x3, and optimizer Adam with epoch 200. The accuracy of the image classification process is 80%. The non-demented label predicts that the uncertainty is 0.371, and the other uncertainty prediction is 0.002. The ability to anticipate uncertainty enables clinicians to make informed decisions regarding the reliability of the model’s output and the need for additional validation or confirmation.

Motor Bearing Failure Identification Using Multiple Long Short-Term Memory Training Strategies

In the context of condition-based maintenance of rotating machines in manufacturing systems, the early diagnosis of possible faults related to rolling elements of the bearing is mainly based on techniques from artificial intelligence, namely, Machine Learning (ML) and Deep Learning (DL). Approaches based on using Deep Learning methods have been the most coveted in recent years. Among a variety of models, the type of architecture known as Long-Short-Term Memory (LSTM) of Recurrent Neural Network (RNN) has both the ability to capture long-term dependencies and to adapt to sequential data modeling. It is therefore able to work on data without any preprocessing. This paper studies using four types of LSTM networks to diagnose bearing faults in a classification approach. It aims to intervene on both the input parameters and the network architecture, to achieve high performance. The proposed method is carried out in two different ways. In the first case, the data inputs are raw frames of vibration signals. However, in the second case, the network inputs are pre-computed time-frequency features. The results clearly showed that LSTMs are more accurate with the latter.

MVD-HG: multigranularity smart contract vulnerability detection method based on heterogeneous graphs

Smart contracts have significant losses due to various types of vulnerabilities. However, traditional vulnerability detection methods rely extensively on expert rules, resulting in low detection accuracy and poor adaptability to novel attacks. To address these problems, in this paper, deep learning methods are combined with smart contract vulnerability code detection approaches. syntax trees (ASTs), which are special isomorphic graph structures, are an important bridge between source code and graph neural networks. By learning the AST, the model can understand the semantics of the source code. Moreover, graph neural networks have an increasing ability to address complex heterogeneous graphs. Therefore, control flow graphs are fused with data flow graphs on the basis of the ASTs to build heterogeneous graphs with richer code semantics. Furthermore, multigranularity analysis of the vulnerability detection results is performed, including coarse-grained contract-level vulnerability detection and fine-grained line-level vulnerability detection. Through this multigranularity detection approach, vulnerabilities in contracts can be identified and analysed more comprehensively, providing a richer perspective and more solutions for vulnerability detection. The experimental results show that the proposed multigranularity vulnerability detection method based on heterogeneous graphs (MVD-HG) improves both the accuracy and range of the detected vulnerability types in contract-level vulnerability detection tasks; moreover, in the line-level vulnerability detection task, the MVD-HG model achieves significant results and addresses the shortcomings of existing methods. In addition, based on code generation methods used in related fields, a data enhancement method based on the source code is developed, which effectively expands the experimental dataset to address the reduced credibility of the results due to insufficient amounts of data.

PointNet + + Based Concealed Object Classification Utilizing an FMCW Millimeter-Wave Radar

AbstractIn the field of millimeter-wave (MMW) imaging, the integration of artificial intelligence (AI) has emerged as a crucial solution for addressing automation challenges. In this study, concealed object classification was successfully achieved on point cloud data from MMW radar high-precision imaging using the PointNet + + deep learning method. The utilized dataset comprises point cloud data generated through the transformation of 3D models and reconstruction of physical objects with an accuracy of less than 1 mm via MMW radar scanning. Classification accuracy was significantly improved by introducing data enhancement techniques, including the generation of homologous data and optimization of sampling points. After several evaluations, 300 epochs of training were conducted using 8192 sampling points, the results showed an accuracy of 0.998 for the training dataset and 0.996 for the test dataset. Moreover, evaluations of samples not included in the original dataset as well as multi-surface scans of concealed objects within the cardboard both resulted in correct predictions, which further validates the effectiveness and reliability of the study and demonstrates the potential of AI applied to MMW imaging.

Parking Slot Detection and Vacancy Check Based on Deep Learning Method

In Bangladesh, traffic is a huge problem, mostly in urban areas like Dhaka. Parking at the roadside is a major problem for traffic in Bangladesh. Private transport user waste most of their time in their daily life searching for a parking slot. Therefore, there is a crucial need to develop a system to help people find vacant parking spots. Systems for managing parking spaces, such as those that identify empty spaces, can help minimize traffic and energy waste in big cities. Since visual approaches may use security cameras installed already in several parking lots, they are an affordable alternative to other methods of vacancy detection. Based on the deep learning model, we have made a cost-effective system and will apply data from that camera. After applying different deep-learning models, we implemented YOLOv7, Mask-RCNN, and YOLOv5. Class loss, Box loss, instances, mAp, and Object loss are all generated by the model. Among all models, YOLOv5 has the highest mAp, 98%. In the future, we will work on providing access to a registered user in the parking garage. This will increase the dataset and obtain higher accuracy.

Multi-Layer Perceptron and Radial Basis Function Networks in Predictive Modeling of Churn for Mobile Telecommunications Based on Usage Patterns

Customer retention is a key priority for mobile telecommunications companies, as acquiring new customers is significantly more costly than retaining existing ones. A major challenge in this field is predicting customer churn—users discontinuing services. Traditional predictive models such as rule-based systems often struggle with the complex, non-linear nature of customer behavior. To address this, we propose the use of deep learning techniques, specifically multi-layer perceptron (MLP) and radial basis function (RBF) networks, to improve the accuracy of churn predictions. However, while neural networks excel in predictive performance, they are often criticized for being “black-box” models, lacking interpretability. A real-world data set is considered, which originally contained information about 15,000 randomly selected clients. Various network structures and configurations are analyzed. The obtained results are compared with results generated using fuzzy rule-based and rough-set rule-based systems. The MLP model achieved an almost perfect accuracy of 0.999 with an F-measure of 0.989, outperforming traditional methods such as fuzzy rule-based and rough-set systems. Although the RBF model slightly lagged in accuracy, it demonstrated a superior recall of 0.993, indicating better identification of potential churners. These results demonstrate that neural network models significantly enhance predictive performance in churn modeling. The interpretability of the model is also discussed since it bears significance in real applications. Our contribution lies in showing that deep learning methods significantly enhance churn prediction accuracy, though the challenge of model interpretability remains a critical area for future work.

Mutual information-based radiomic feature selection with SHAP explainability for breast cancer diagnosis

Breast cancer is a prevalent concern for women globally, with misdiagnosis potentially leading to detrimental outcomes. Early detection is crucial, often reliant on medical imaging analysis. Digital Breast Tomosynthesis (DBT) is a promising modality, addressing limitations of traditional mammograms. However, diagnosing breast cancer involves subjective visual examination, leading to inaccuracies. Radiomics, applied in various imaging modalities such as MRI, and digital mammography, remains underutilized in DBT. This study introduces a Mutual Information-based Radiomic Feature Selection (MIRFS) framework for DBT breast cancer evaluation followed by SHAP explanations. Selected features were assessed using machine learning algorithms, with Random Forest achieving 92% accuracy in lesion classification. MIRFS demonstrates significant performance improvements, addressing subjectivity and enhancing diagnostic accuracy through explainability. SHAP methodology elucidates feature importance, aiding model interpretation. Compared to deep learning methods, MIRFS outperforms both deep learning and existing machine learning approaches, promising advancements in breast cancer diagnosis and treatment.

Modeling the CO2 Emissions of Turkey Dependent on Various Parameters Employing ARIMAX and Deep Learning Methods

Modeling the CO2 Emissions of Turkey Dependent on Various Parameters Employing ARIMAX and Deep Learning Methods

A Critical Review on Sentiment Analysis Based on Deep Learning Techniques

Sentiment analysis, a vital task in natural language processing, has evolved significantly with the adoption of deep learning techniques. This review critically examines the current state of sentiment analysis based on deep learning methods, focusing on their performance, scalability, and challenges. We explore key deep learning models such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), Long Short-Term Memory (LSTM), and attention mechanisms. These models have shown remarkable improvements in sentiment prediction accuracy compared to traditional machine learning approaches. However, issues like data scarcity, interpretability, and computational complexity remain challenging. This review provides insights into existing solutions, evaluates emerging trends, and outlines future directions to enhance deep learning applications in sentiment analysis across diverse domains.

State of the Art and Potentialities of Graph-level Learning

Graphs have a superior ability to represent relational data, such as chemical compounds, proteins, and social networks. Hence, graph-level learning, which takes a set of graphs as input, has been applied to many tasks, including comparison, regression, classification, and more. Traditional approaches to learning a set of graphs heavily rely on hand-crafted features, such as substructures. While these methods benefit from good interpretability, they often suffer from computational bottlenecks, as they cannot skirt the graph isomorphism problem. Conversely, deep learning has helped graph-level learning adapt to the growing scale of graphs by extracting features automatically and encoding graphs into low-dimensional representations. As a result, these deep graph learning methods have been responsible for many successes. Yet, no comprehensive survey reviews graph-level learning starting with traditional learning and moving through to the deep learning approaches. This article fills this gap and frames the representative algorithms into a systematic taxonomy covering traditional learning, graph-level deep neural networks, graph-level graph neural networks, and graph pooling. In addition, the evolution and interaction between methods from these four branches within their developments are examined to provide an in-depth analysis. This is followed by a brief review of the benchmark datasets, evaluation metrics, and common downstream applications. Finally, the survey concludes with an in-depth discussion of 12 current and future directions in this booming field.

Comparison of deep-learning multimodality data fusion strategies in mandibular osteoradionecrosis NTCP modelling using clinical variables and radiation dose distribution volumes

Objective.Normal tissue complication probability (NTCP) modelling is rapidly embracing deep learning (DL) methods, acknowledging the importance of spatial dose information. Finding effective ways to combine information from radiation dose distribution maps (dosiomics) and clinical data involves technical challenges and requires domain knowledge. We propose different multi-modality data fusion strategies to facilitate future DL-based NTCP studies.Approach.Early, joint and late DL multi-modality fusion strategies were compared using clinical and mandibular radiation dose distribution volumes. These were contrasted with single-modality models: a random forest trained on non-image data (clinical, demographic and dose-volume metrics) and a 3D DenseNet-40 trained on image data (mandibular dose distribution maps). The study involved a matched cohort of 92 osteoradionecrosis cases and 92 controls from a single institution.Main results.The late fusion model exhibited superior discrimination and calibration performance, while the join fusion achieved a more balanced distribution of the predicted probabilities. Discrimination performance did not significantly differ between strategies. Late fusion, though less technically complex, lacks crucial inter-modality interactions for NTCP modelling. In contrast, joint fusion, despite its complexity, resulted in a single network training process which included intra- and inter-modality interactions in its model parameter optimisation.Significance.This study is a pioneering effort in comparing different strategies for including image data into DL-based NTCP models in combination with lower dimensional data such as clinical variables. The discrimination performance of such multi-modality NTCP models and the choice of fusion strategy will depend on the distribution and quality of both types of data. Multiple data fusion strategies should be compared and reported in multi-modality NTCP modelling using DL.

Deep learning-based detection of affected body parts in Parkinson’s disease and freezing of gait using time-series imaging

We proposed a deep learning method using a convolutional neural network on time-series (TS) images to detect and differentiate affected body parts in people with Parkinson’s disease (PD) and freezing of gait (FOG) during 360° turning tasks. The 360° turning task was performed by 90 participants (60 people with PD [30 freezers and 30 nonfreezers] and 30 age-matched older adults (controls) at their preferred speed. The position and acceleration underwent preprocessing. The analysis was expanded from temporal to visual data using TS imaging methods. According to the PD vs. controls classification, the right lower third of the lateral shank (RTIB) on the least affected side (LAS) and the right calcaneus (RHEE) on the LAS were the most relevant body segments in the position and acceleration TS images. The RHEE marker exhibited the highest accuracy in the acceleration TS images. The identified markers for the classification of freezers vs. nonfreezers vs. controls were the left lateral humeral epicondyle (LELB) on the more affected side and the left posterior superior iliac spine (LPSI). The LPSI marker in the acceleration TS images displayed the highest accuracy. This approach could be a useful supplementary tool for determining PD severity and FOG.

Assessment of Emphysema on X-ray Equivalent Dose Photon-Counting Detector CT

Objectives The aim of this study was to evaluate the feasibility and efficacy of visual scoring, low-attenuation volume (LAV), and deep learning methods for estimating emphysema extent in x-ray dose photon-counting detector computed tomography (PCD-CT), aiming to explore future dose reduction potentials. Methods One hundred one prospectively enrolled patients underwent noncontrast low- and chest x-ray dose CT scans in the same study using PCD-CT. Overall image quality, sharpness, and noise, as well as visual emphysema pattern (no, trace, mild, moderate, confluent, and advanced destructive emphysema; as defined by the Fleischner Society), were independently assessed by 2 experienced radiologists for low- and x-ray dose images, followed by an expert consensus read. In the second step, automated emphysema quantification was performed using an established LAV algorithm with a threshold of −950 HU and a commercially available deep learning model for automated emphysema quantification. Automated estimations of emphysema extent were converted and compared with visual scoring ratings. Results X-ray dose scans exhibited a significantly lower computed tomography dose index than low-dose scans (low-dose: 0.66 ± 0.16 mGy, x-ray dose: 0.11 ± 0.03 mGy, P < 0.001). Interreader agreement between low- and x-ray dose for visual emphysema scoring was excellent (κ = 0.83). Visual emphysema scoring consensus showed good agreement between low-dose and x-ray dose scans (κ = 0.70), with significant and strong correlation (Spearman ρ = 0.79). Although trace emphysema was underestimated in x-ray dose scans, there was no significant difference in the detection of higher-grade (mild to advanced destructive) emphysema (P = 0.125) between the 2 scan doses. Although predicted emphysema volumes on x-ray dose scans for the LAV method showed strong and the deep learning model excellent significant correlations with predictions on low-dose scans, both methods significantly overestimated emphysema volumes on lower quality scans (P < 0.001), with the deep learning model being more robust. Further, deep learning emphysema severity estimations showed higher agreement (κ = 0.65) and correlation (Spearman ρ = 0.64) with visual scoring for low-dose scans than LAV predictions (κ = 0.48, Spearman ρ = 0.45). Conclusions The severity of emphysema can be reliably estimated using visual scoring on CT scans performed with x-ray equivalent doses on a PCD-CT. A deep learning algorithm demonstrated good agreement and strong correlation with the visual scoring method on low-dose scans. However, both the deep learning and LAV algorithms overestimated emphysema extent on x-ray dose scans. Nonetheless, x-ray equivalent radiation dose scans may revolutionize the detection and monitoring of disease in chronic obstructive pulmonary disease patients.

A deep learning-based dose calculation method for volumetric modulated arc therapy

BackgroundVolumetric modulated arc therapy (VMAT) planning optimization involves iterative adjustment of numerous parameters, and hence requires repeatedly dose recalculation. In this study, we used the deep learning method to develop a fast and accurate dose calculation method for VMAT.MethodsThe classical 3D UNet was adopted and trained to learn the physics principle of dose calculation. The inputs included the projected fluence map (FM), computed tomography (CT) images, the radiological depth and the source-to-voxel distance (SVD). The projected FM was generated by projecting the accumulated FM between two consecutive control points (CPs) onto the patient’s anatomy. The accumulated FM was calculated by simulating the movement of the multi-leaf collimator (MLC) from one CP to the next. The dose, calculated by the treatment planning system (TPS), was used as ground truth. 51 head and neck VMAT plans were used, with 43, 1 and 7 cases as training, validation, and testing datasets, respectively. Correspondingly, 7182, 180 and 1260 CP samples were included in the training, validation, and testing datasets.ResultsThis presented method was evaluated by comparing the derived dose distribution to the TPS calculated dose distribution. The dose profiles coincided for both the single CP and the entire plan (summation of all CPs). But the network derived dose was smoother than the TPS calculated dose. Gamma analysis was performed between the network derived dose and the TPS calculated dose. The average gamma pass rate was 96.56%, 98.75%, 98.03% and 99.30% under the criteria of 2% (tolerance) -2 mm (distance to agreement, DTA). 2%-3 mm, 3%-2 mm and 3%-3 mm. No significant difference was observed on the critical indices including the max, mean dose, and the relative volume covered by the 2000 cGy, 4000 cGy and the prescription dose. For one CP, the average computational time of the network and TPS was 0.09s and 0.53s. And for one patient, the average time was 16.51s and 95.60s.ConclusionThe dose distribution derived by the network showed good agreement with the TPS calculated dose distribution. The computational time was reduced to approximate one-sixth of its original duration. Therefore the presented deep learning-based dose calculation method has the potential to be used for planning optimization.

A deep learning method for beat-level risk analysis and interpretation of atrial fibrillation patients during sinus rhythm

Atrial Fibrillation (AF) is a common cardiac arrhythmia. Many AF patients experience complications such as stroke and other cardiovascular issues. Early detection of AF is crucial. The great majority of algorithms can only distinguish “AF rhythm in AF patients” from “sinus rhythm in healthy individuals” . However, AF patients do not always exhibit AF rhythm, most of the time they present with sinus rhythms, and there is also a potential risk of AF in sinus rhythms. How to detect AF from sinus rhythm is a challenge. To address this, this paper proposes a novel artificial intelligence (AI) algorithm to distinguish “sinus rhythm in AF patients” and “sinus rhythm in healthy individuals” in beat-level. We cut 1 s of sinus beats from single-lead Electrocardiogram (ECG) data and fed them to the Net1d model, a deep learning model that processes one-dimensional data, to obtain the risk probability of each beat. Besides, we have also introduced the beat-level risk interpreters, trend risk interpreters, addressing the interpretability issues of deep learning models and the difficulty in explaining AF risk trends. Additionally, the beat-level information fusion decision is presented to enhance model accuracy. The experimental results demonstrate that the average AUC for single beats used as testing data from CPSC 2021 dataset is 0.7314, with an average accuracy of 0.6606 and an F1 score of 0.6470. By employing 150 beats for information fusion decision algorithm, the average AUC can reach 0.7591, while the average accuracy and F1 score improve to 0.6887 and 0.6749. Compared to previous segment-level algorithms, we utilized beats as input, reducing data dimensionality and making the model more lightweight, facilitating deployment on portable medical devices. Furthermore, we draw new and interesting findings through average beat analysis and subgroup analysis, considering varying risk levels. Our code is publicly available at https://github.com/leijsen/ECGBeat4AFSinus.

Spatial-temporal deep learning method for solving data-driven multi-echelon stochastic lot sizing problems with intermediate demands

ABSTRACT For the production processes and inventory decisions of a multichannel supply chain, a multi-echelon stochastic lot sizing problem with intermediate demands is extended to a constrained problem that considers the production (inventory) capacity constraint and to multi-item problems that consider feature and item correlations. Three spatial-temporal deep learning methods combining convolutional, recurrent and graph convolutional neural networks are proposed to transform the extended problems as spatial-temporal sequential prediction problems and learn data-driven decisions. Real-world data are used in the computational experiments, and the results show that considering both temporal and spatial features and/or feature correlations can provide improved prediction performance.

A Performance Comparison Study on Climate Prediction in Weifang City Using Different Deep Learning Models

Climate change affects the water cycle, water resource management, and sustainable socio-economic development. In order to accurately predict climate change in Weifang City, China, this study utilizes multiple data-driven deep learning models. The climate data for 73 years include monthly average air temperature (MAAT), monthly average minimum air temperature (MAMINAT), monthly average maximum air temperature (MAMAXAT), and monthly total precipitation (MP). The different deep learning models include artificial neural network (ANN), recurrent NN (RNN), gate recurrent unit (GRU), long short-term memory neural network (LSTM), deep convolutional NN (CNN), hybrid CNN-GRU, hybrid CNN-LSTM, and hybrid CNN-LSTM-GRU. The CNN-LSTM-GRU for MAAT prediction is the best-performing model compared to other deep learning models with the highest correlation coefficient (R = 0.9879) and lowest root mean square error (RMSE = 1.5347) and mean absolute error (MAE = 1.1830). These results indicate that The hybrid CNN-LSTM-GRU method is a suitable climate prediction model. This deep learning method can also be used for surface water modeling. Climate prediction will help with flood control and water resource management.

A systematic scrutiny of artificial intelligence-based air pollution prediction techniques, challenges, and viable solutions

Air is an essential human necessity, and inhaling filthy air poses a significant health risk. One of the most severe hazards to people’s health is air pollution, and appropriate precautions should be taken to monitor and anticipate its quality in advance. Among all the countries, the air quality in India is decreasing daily, which is a matter of concern to the health department. Many studies use machine learning and Deep learning methods to predict atmospheric pollutant levels, prioritizing accuracy over interpretability. Many research studies confuse researchers and readers about how to proceed with further research. This paper aims to give every detail of the considered air pollutants and brief about the techniques used, their advantages, and challenges faced during pollutant prediction, which leads to a better understanding of the techniques before starting any research related to air pollutant prediction. This paper has given numerous prospective questions on air pollution that piqued the study’s interest. This study discussed various machine and deep learning methods and optimization techniques. Despite all the discussed machine learning and deep learning techniques, the paper concluded that more datasets, better learning techniques, and a variety of suggestions would enhance interpretability while maintaining high accuracy for air pollution prediction. The purpose of this review is also to reveal how a family of neural network algorithms has helped researchers across the globe to predict air pollutant(s).