Journal of Big Data最新文献

This study introduces a novel deep learning-based approach for classifying poultry audio signals, incorporating a custom Burn Layer to enhance model robustness. The methodology integrates digital audio signal processing, convolutional neural networks (CNNs), and the innovative Burn Layer, which injects controlled random noise during training to reinforce the model's resilience to input signal variations. The proposed architecture is streamlined, with convolutional blocks, densely connected layers, dropout, and an additional Burn Layer to fortify robustness. The model demonstrates efficiency by reducing trainable parameters to 191,235, compared to traditional architectures with over 1.7 million parameters. The proposed model utilizes a Burn Layer with burn intensity as a parameter and an Adamax optimizer to optimize and address the overfitting problem. Thorough evaluation using six standard classification metrics showcases the model's superior performance, achieving exceptional sensitivity (96.77%), specificity (100.00%), precision (100.00%), negative predictive value (NPV) (95.00%), accuracy (98.55%), F1 score (98.36%), and Matthew’s correlation coefficient (MCC) (95.88%). This research contributes valuable insights into the fields of audio signal processing, animal health monitoring, and robust deep-learning classification systems. The proposed model presents a systematic approach for developing and evaluating a deep learning-based poultry audio classification system. It processes raw audio data and labels to generate digital representations, utilizes a Burn Layer for training variability, and constructs a CNN model with convolutional blocks, pooling, and dense layers. The model is optimized using the Adamax algorithm and trained with data augmentation and early-stopping techniques. Rigorous assessment on a test dataset using standard metrics demonstrates the model's robustness and efficiency, with the potential to significantly advance animal health monitoring and disease detection through audio signal analysis.

In late 2023, the United Nations conference on climate change (COP28), which was held in Dubai, encouraged a quick move from fossil fuels to renewable energy. Solar energy is one of the most promising forms of energy that is both sustainable and renewable. Generally, photovoltaic systems transform solar irradiance into electricity. Unfortunately, instability and intermittency in solar radiation can lead to interruptions in electricity production. The accurate forecasting of solar irradiance guarantees sustainable power production even when solar irradiance is not present. Batteries can store solar energy to be used during periods of solar absence. Additionally, deterministic models take into account the specification of technical PV systems and may be not accurate for low solar irradiance. This paper presents a comparative study for the most common Deep Learning (DL) and Machine Learning (ML) algorithms employed for short-term solar irradiance forecasting. The dataset was gathered in Islamabad during a five-year period, from 2015 to 2019, at hourly intervals with accurate meteorological sensors. Furthermore, the Grid Search Cross Validation (GSCV) with five folds is introduced to ML and DL models for optimizing the hyperparameters of these models. Several performance metrics are used to assess the algorithms, such as the Adjusted R² score, Normalized Root Mean Square Error (NRMSE), Mean Absolute Deviation (MAD), Mean Absolute Error (MAE) and Mean Square Error (MSE). The statistical analysis shows that CNN-LSTM outperforms its counterparts of nine well-known DL models with Adjusted R² score value of 0.984. For ML algorithms, gradient boosting regression is an effective forecasting method with Adjusted R² score value of 0.962, beating its rivals of six ML models. Furthermore, SHAP and LIME are examples of explainable Artificial Intelligence (XAI) utilized for understanding the reasons behind the obtained results.

The frequent usage of computer networks and the Internet has made computer networks vulnerable to numerous attacks, highlighting the critical need to enhance the precision of security mechanisms. One of the most essential measures to safeguard networking resources and infrastructures is an intrusion detection system (IDS). IDSs are widely used to detect, identify, and track malicious threats. Although various machine learning algorithms have been used successfully in IDSs, they are still suffering from low prediction performances. One reason behind the low accuracy of IDSs is that existing network traffic datasets have high computational complexities that are mainly caused by redundant, incomplete, and irrelevant features. Furthermore, standalone classifiers exhibit restricted classification performance and typically fail to produce satisfactory outcomes when dealing with imbalanced, multi-category traffic data. To address these issues, we propose an efficient intrusion detection model, which is based on hybrid feature selection and stack ensemble learning. Our hybrid feature selection method, called MI-Boruta, combines mutual information (MI) as a filter method and the Boruta algorithm as a wrapper method to determine optimal features from our datasets. Then, we apply stacked ensemble learning by using random forest (RF), Catboost, and XGBoost algorithms as base learners with multilayer perceptron (MLP) as meta-learner. We test our intrusion detection model on two widely recognized benchmark datasets, namely UNSW-NB15 and CICIDS2017. We show that our proposed IDS outperforms existing IDSs in almost all performance criteria, including accuracy, recall, precision, F1-Score, false positive rate, true positive rate, and error rate.

Background

The tumor microenvironment (TME) provides a region for intricate interactions within or between immune and non-immune cells. We aimed to reveal the tissue architecture and comprehensive landscape of cells within the TME of colorectal cancer (CRC).

Methods

Fresh frozen invasive adenocarcinoma of the large intestine tissue from 10× Genomics Datasets was obtained from BioIVT Asterand. The integration of microarray-based spatial transcriptomics (ST) and RNA sequencing (RNA-seq) was applied to characterize gene expression and cell landscape within the TME of CRC tissue architecture. Multiple R packages and deconvolution algorithms including MCPcounter, XCELL, EPIC, and ESTIMATE methods were performed for further immune distribution analysis.

Results

The subpopulations of immune and non-immune cells within the TME of the CRC tissue architecture were appropriately annotated. According to ST and RNA-seq analyses, a heterogeneous spatial atlas of gene distribution and cell landscape was comprehensively characterized. We distinguished between the cancer and stromal regions of CRC tissues. As expected, epithelial cells were located in the cancerous region, whereas fibroblasts were mainly located in the stroma. In addition, the fibroblasts were further subdivided into two subgroups (F1 and F2) according to the differentially expressed genes (DEGs), which were mainly enriched in pathways including hallmark-oxidative-phosphorylation, hallmark-e2f-targets and hallmark-unfolded-protein-response. Furthermore, the top 5 DEGs, SPP1, CXCL10, APOE, APOC1, and LYZ, were found to be closely related to immunoregulation of the TME, methylation, and survival of CRC patients.

Conclusions

This study characterized the heterogeneous spatial landscape of various cell subtypes within the TME of the tissue architecture. The TME-related roles of fibroblast subsets addressed the potential crosstalk among diverse cells.

Diabetic retinopathy (DR) is the most prevalent cause of preventable vision loss worldwide, imposing a significant economic and medical burden on society today, of which early identification is the cornerstones of the management. The diagnosis and severity grading of DR rely on scales based on clinical visualized features, but lack detailed quantitative parameters. Retinal non-perfusion area (NPA) is a pathogenic characteristic of DR that symbolizes retinal hypoxia conditions, and was found to be intimately associated with disease progression, prognosis, and management. However, the practical value of NPA is constrained since it appears on fundus fluorescein angiography (FFA) as distributed, irregularly shaped, darker plaques that are challenging to measure manually. In this study, we propose a deep learning-based method, NPA-Net, for accurate and automatic segmentation of NPAs from FFA images acquired in clinical practice. NPA-Net uses the U-net structure as the basic backbone, which has an encoder-decoder model structure. To enhance the recognition performance of the model for NPA, we adaptively incorporate multi-scale features and contextual information in feature learning and design three modules: Adaptive Encoder Feature Fusion (AEFF) module, Multilayer Deep Supervised Loss, and Atrous Spatial Pyramid Pooling (ASPP) module, which enhance the recognition ability of the model for NPAs of different sizes from different perspectives. We conducted extensive experiments on a clinical dataset with 163 eyes with NPAs manually annotated by ophthalmologists, and NPA-Net achieved better segmentation performance compared to other existing methods with an area under the receiver operating characteristic curve (AUC) of 0.9752, accuracy of 0.9431, sensitivity of 0.8794, specificity of 0.9459, IOU of 0.3876 and Dice of 0.5686. This new automatic segmentation model is useful for identifying NPA in clinical practice, generating quantitative parameters that can be useful for further research as well as guiding DR detection, grading severity, treatment planning, and prognosis.

Background

The long-term impacts of COVID-19 on human health are a major concern, yet comprehensive evaluations of its effects on various health conditions are lacking.

Methods

This study aims to evaluate the role of various diseases in relation to COVID-19 by analyzing genetic data from a large-scale population over 2,000,000 individuals. A bidirectional two-sample Mendelian randomization approach was used, with exposures including COVID-19 susceptibility, hospitalization, and severity, and outcomes encompassing 86 different diseases or traits. A reverse Mendelian randomization analysis was performed to assess the impact of these diseases on COVID-19.

Results

Our analysis identified causal relationships between COVID-19 susceptibility and several conditions, including breast cancer (OR = 1.0073, 95% CI = 1.0032–1.0114, p = 5 × 10 − 4), ER + breast cancer (OR = 0.5252, 95% CI = 0.3589–0.7685, p = 9 × 10 − 4), and heart failure (OR = 1.0026, 95% CI = 1.001–1.0042, p = 0.002). COVID-19 hospitalization was causally linked to heart failure (OR = 1.0017, 95% CI = 1.0006–1.0028, p = 0.002) and Alzheimer’s disease (OR = 1.5092, 95% CI = 1.1942–1.9072, p = 0.0006). COVID-19 severity had causal effects on primary biliary cirrhosis (OR = 2.6333, 95% CI = 1.8274–3.7948, p = 2.059 × 10 − 7), celiac disease (OR = 0.0708, 95% CI = 0.0538–0.0932, p = 9.438 × 10–80), and Alzheimer’s disease (OR = 1.5092, 95% CI = 1.1942–1.9072, p = 0.0006). Reverse MR analysis indicated that rheumatoid arthritis, diabetic nephropathy, multiple sclerosis, and total testosterone (female) influence COVID-19 outcomes. We assessed heterogeneity and horizontal pleiotropy to ensure result reliability and employed the Steiger directionality test to confirm the direction of causality.

Conclusions

This study provides a comprehensive analysis of the causal relationships between COVID-19 and diverse health conditions. Our findings highlight the long-term impacts of COVID-19 on human health, emphasizing the need for continuous monitoring and targeted interventions for affected individuals. Future research should explore these relationships to develop comprehensive healthcare strategies.

The image object recognition and detection technology are widely used in many scenarios. In recent years, big data has become increasingly abundant, and big data-driven artificial intelligence models have attracted more and more attention. Evolutionary computation has also provided a powerful driving force for the optimization and improvement of deep learning models. In this paper, we propose an image object detection method based on self-supervised and data-driven learning. Differ from other methods, our approach stands out due to its innovative use of multispectral data fusion and evolutionary computation for model optimization. Specifically, our method uniquely combines visible light images and infrared images to detect and identify image targets. Firstly, we utilize a self-supervised learning method and the AutoEncoder model to perform high-dimensional feature extraction on the two types of images. Secondly, we fuse the extracted features from the visible light and infrared images to detect and identify objects. Thirdly, we introduce a model parameter optimization method using evolutionary learning algorithms to enhance model performance. Validation on public datasets shows that our method achieves comparable or superior performance to existing methods.

This article presents an investment recommender system based on an Adaptive Neuro-Fuzzy Inference System (ANFIS) and pre-trained weights from a Multimodal Neural Network (MNN). The model is designed to support the investment process for the customers and takes into consideration seven factors to implement the proposed investment system model through the customer or potential investor data set. The system takes input from a web-based questionnaire that collects data on investors' preferences and investment goals. The data is then preprocessed and clustered using ETL tools, JMP, MATLAB, and Python. The ANFIS-based recommender system is designed with three inputs and one output and trained using a hybrid approach over three epochs with 188 data pairs and 18 fuzzy rules. The system's performance is evaluated using metrics such as RMSE, accuracy, precision, recall, and F1-score. The system is also designed to incorporate expert feedback and opinions from investors to customize and improve investment recommendations. The article concludes that the proposed ANFIS-based investment recommender system is effective and accurate in generating investment recommendations that meet investors' preferences and goals.

Graphical abstract

Psychiatric disorders are severe health challenges that exert a heavy public burden. Air pollution has been widely reported as related to psychiatric disorder risk, but their casual association and pathological mechanism remained unclear. Herein, we systematically investigated the large genome-wide association studies (6 cohorts with 1,357,645 samples), single-cell RNA (26 samples with 157,488 cells), and bulk-RNAseq (1595 samples) datasets to reveal the genetic causality and biological link between four air pollutants and nine psychiatric disorders. As a result, we identified ten positive genetic correlations between air pollution and psychiatric disorders. Besides, PM2.5 and NO₂ presented significant causal effects on schizophrenia risk which was robust with adjustment of potential confounders. Besides, transcriptome-wide association studies identified the shared genes between PM2.5/NO2 and schizophrenia. We then discovered a schizophrenia-derived inhibitory neuron subtype with highly expressed shared genes and abnormal synaptic and metabolic pathways by scRNA analyses and confirmed their abnormal level and correlations with the shared genes in schizophrenia patients in a large RNA-seq cohort. Comprehensively, we discovered robust genetic causality between PM2.5, NO₂, and schizophrenia and identified an abnormal inhibitory neuron subtype that links schizophrenia pathology and PM2.5/NO2 exposure. These discoveries highlight the schizophrenia risk under air pollutants exposure and provide novel mechanical insights into schizophrenia pathology, contributing to pollutant-related schizophrenia risk control and therapeutic strategies development.

Graphical Abstract

Data has evolved into one of the principal resources for contemporary businesses. Moreover, corporations have undergone digitalization; consequently, their supply chains generate substantial amounts of data. The theoretical framework of this investigation was built on novel concepts like big data analytics—artificial intelligence (BDA-AI) and supply chain ambidexterity’s (SCA) direct impacts on sustainable supply chain management (SSCM) and indirect impacts on sustainable innovation ambidexterity (SIA) and environmental performance (EP). This study selected employees of manufacturing industries as respondents for environmental performance, sustainable supply chain management, big data analytics, artificial intelligence, and supply chain ambidexterity. The results from this study show that BDA-AI and SCA significantly affect SSCM. SSCM has significant associations with SIA and EP. Finally, SIA has a significant impact on EP. According to the results indicating the indirect impacts, BDA-AI has significant indirect relationships with SIA and EP by having SSCM as the mediating variable. Furthermore, SCA has significant indirect associations with SIA and EP, with SSCM as the mediating variable. Additionally, both BDA-AI and SCA have significant indirect associations with EP, while SIA and SSCM are mediating variables. Finally, SSCM has an indirect association with EP while having SIA as a mediating variable. The findings of this paper provide several theoretical contributions to the research in sustainability and big data analytics artificial intelligence field. Furthermore, based on the suggested framework, this study offers a number of practical implications for decision-makers to improve significantly in the supply chain and BDA-AI. For instance, this paper provides significant insight for logistics and supply chain managers, supporting them in implementing BDA-AI solutions to help SSCM and enhance EP.