Integrating Machine Learning Research Articles

Integrating machine learning for enhanced wildfire severity prediction: A study in the Upper Colorado River basin

Wildfires pose significant threats worldwide, requiring accurate prediction for mitigation. This study uses machine learning techniques to forecast wildfire severity in the Upper Colorado River basin. Datasets from 1984 to 2019 and key indicators like weather conditions and land use were employed. Random Forest outperformed Artificial Neural Network, achieving 72 % accuracy. Influential predictors include air temperature, vapor pressure deficit, NDVI, and fuel moisture. Solar radiation, SPEI, precipitation, and evapotranspiration also contribute significantly. Validation against actual severities from 2016 to 2019 showed mean prediction errors of 11.2 %, affirming the model's reliability. These results highlight the efficacy of machine learning in understanding wildfire severity, especially in vulnerable regions.

Integrated machine learning survival framework to decipher diverse cell death patterns for predicting prognosis in lung adenocarcinoma.

Various forms of programmed cell death (PCD) collectively regulate the occurrence, development and metastasis of tumors. Nevertheless, a comprehensive analysis of the diverse types of PCD in lung adenocarcinoma (LUAD) is currently lacking. The study encompassed a total of 1481 genes associated with the regulation of 13 distinct PCD patterns. Ten machine learning algorithms were amalgamated into 101 combinations, from which the optimal algorithm was chosen to formulate an artificial intelligence-derived prognostic signature based on the average C-index across four multicenter cohorts. The established optimal cell death index (CDI) model emerged as an independent risk factor for overall survival, demonstrating robust and consistent performance. Notably, CDI exhibited significantly higher accuracy compared to traditional clinical variables and molecular features. It exhibited superior performance than other published models. By integrating CDI with relevant clinical features, a nomogram with excellent predictive performance was developed. LUAD patients with low CDI score had a higher immune modulators, TIDE scores and immune scores, indicating a better immunotherapy benefit. More importantly, we found that the regulation of antigen presentation is the crucial mechanism of PCD. SCG2 is a key molecule that inhibits the malignant progression of LUAD. CDI holds great potential as a robust and promising tool for enhancing clinical outcomes in patients with LUAD.

A greater negative impact of future climate change on vegetation in Central Asia: Evidence from trajectory/pattern analysis

In the context of global warming, vegetation changes exhibit various patterns, yet previous studies have focused primarily on monotonic changes, often overlooking the complexity and diversity of multiple change processes. Therefore, it is crucial to further explore vegetation dynamics and diverse change trajectories in this region under future climate scenarios to obtain a more comprehensive understanding of local ecosystem evolution. In this study, we established an integrated machine learning prediction framework and a vegetation change trajectory recognition framework to predict the dynamics of vegetation in Central Asia under future climate change scenarios and identify its change trajectories, thus revealing the potential impacts of future climate change on vegetation in the region. The findings suggest that various future climate scenarios will negatively affect most vegetation in Central Asia, with vegetation change intensity increasing with increasing emission trajectories. Analyses of different time scales and trend variations consistently revealed more pronounced downward trends. Vegetation change trajectory analysis revealed that most vegetation has undergone nonlinear and dramatic changes, with negative changes outnumbering positive changes and curve changes outnumbering abrupt changes. Under the highest emission scenario (SSP5-8.5), the abrupt vegetation changes and curve changes are 1.7 times and 1.3 times greater, respectively, than those under the SSP1-2.6 scenario. When transitioning from lower emission pathways (SSP1-2.6, SSP2-4.5) to higher emission pathways (SSP3-7.0, SSP5-8.5), the vegetation change trajectories shift from neutral and negative curve changes to abrupt negative changes. Across climate scenarios, the key climate factors influencing vegetation changes are mostly evapotranspiration and soil moisture, with temperature and relative humidity exerting relatively minor effects. Our study reveals the negative response of vegetation in Central Asia to climate change from the perspective of vegetation dynamics and change trajectories, providing a scientific basis for the development of effective ecological protection and climate adaptation strategies.

Advances in Cybersecurity and AI: Integrating Machine Learning, IoT, and Smart Systems for Resilience and Innovation Across Domains

Artificial Intelligence (AI) is revolutionizing cybersecurity by offering enhanced threat detection, predictive analytics, and automated responses. However, AI also introduces significant challenges, including bias, lack of transparency, and vulnerability to adversarial attacks. This paper examines the dual role of AI in cybersecurity, providing a comprehensive analysis of its benefits and drawbacks, and discussing the future of AI in securing digital environments.

Discrimination of coal geographical origins through HS-GC-IMS assisted with machine learning algorithms in larceny case

The process of globalization and industrialization has resulted in a rise in the theft of coal and other related products, thereby becoming a focal point for forensic science. This situation has engendered an escalated demand for effective detection and monitoring technologies. The precise identification of coal trace evidence presents a challenge with current methods, owing to its minute quantity, fine texture, and intricate composition. In this study, we integrated machine learning with the identification of volatiles to accurately differentiate coal geographical origins through the application of headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS). The topographic distribution of volatiles in coals was visually depicted to elucidate the subtle distinctions through spectra and fingerprint analysis. Additionally, four supervised machine learning algorithms were developed to quantitatively predict the geographical origins of natural coals utilizing the HS-GC-IMS dataset, and these were subsequently compared with unsupervised models. Remarkable volatile compounds were identified through the quantitative analysis and optimal Random Forest model, which offered a rapid readout and achieved an average accuracy of 100 % in coal identification. Our findings indicate that the integration of HS-GC-IMS and machine learning is anticipated to enhance the efficiency and accuracy of coal geographical traceability, thereby providing a foundation for litigation and trials.

Integrating Machine Learning Utility in Tabular Data Synthesizer Training using Loss Function Learning

Machine learning (ML) utility has been the main evaluation metrics for data synthesizers. However, because ML utility cannot be simply calculated, none of the previous synthesizers were trained to reach the same level of ML utility as a training objective. This study aims to integrate ML utility into data synthesizer training using a transformer-based model as a learned loss function. The transformer was trained to estimate ML utility of synthetic datasets, then it’s integrated by backpropagating the difference between estimated and expected value. The integration has significantly improved the average ML utility of LCT-GAN and Realtabformer. The ML utility of LCT-GAN improved by 0.0158 for Contraceptive dataset, 0.031 for Insurance dataset, and 0.0561 for Treatment dataset. The ML utility of Realtabformer improved by 0.02 for Contraceptive dataset and 0.0024 for Insurance dataset. The increase affects the dataset distribution, correlation between features, and privacy, but the direction varies. Correlation coefficients indicate that synthetic data distribution gets closer to real data as ML utility improves. In addition to ML utility integration, this study has also shown that patterns between rows in a dataset can be learned, so better synthesizers can be developed based on them.

Optimization of aerodynamic drag reduction for truck trailer model via machine learning

From advanced fairing designs and spoilers to novel airflow management systems, the quest for drag reduction (DR) in truck trailers is critical for cleaner and more sustainable transport. In this study, a particle swarm optimization (PSO) integrated machine learning approach is developed to optimize the DR performance of a bare truck trailer (BTT) with and without half airfoils and roof fairings. The study aims to reach the optimal truck trailer design with fewer experiments by searching the entire solution space with machine learning methods thanks to the developed approach. For this purpose, the experimental setup was initially designed and the experimental results were obtained. The experimental inputs are Airfoil thickness, Roof Fairing (RF) positions, and Reynolds number, while the experimental output is the drag coefficient. These inputs and outputs were arranged via data preprocessing and then modeled using machine learning methods. Modeling was performed with Artificial Neural Network (ANN) and multiple regression models. In addition to the ANN model setup with appropriate topology, linear and non-linear multiple regression models were established. Comparisons were performed using statistical metrics and it was determined that ANN was the superior model. Using the ANN’s prediction model, the drag coefficients were derived from the whole parameter values of both airfoil thickness and RF positions in the solution space for all Reynolds number and an optimization process was performed by PSO. As a result of the analysis, optimal parameter values were determined as 0.10 and 5.0 mm for the airfoil thickness and RF position. As a result of the experimental validation, it is seen that the BTT_NACA 0010_RF-0.5 model, designed with the optimum parameter values, demonstrated a DR performance of 34.7 %, which is the highest value compared to others. Thus, the heuristic process was experimentally validated and optimized by saving time and cost.

Machine learning approach to predict blood-secretory proteins and potential biomarkers for liver cancer using omics data

Identifying non-invasive blood-based biomarkers is crucial for early detection and monitoring of liver cancer (LC), thereby improving patient outcomes. This study leveraged computational approaches to predict potential blood-based biomarkers for LC. Machine learning (ML) models were developed using selected features from blood-secretory proteins collected from the curated databases. The logistic regression (LR) model demonstrated the optimal performance. Transcriptome analysis across 7 LC cohorts revealed 231 common differentially expressed genes (DEGs). The encoded proteins of these DEGs were compared with the ML dataset, revealing 29 proteins overlapping with the blood-secretory dataset. The LR model also predicted 29 additional proteins as blood-secretory with the remaining protein-coding genes. As a result, 58 potential blood-secretory proteins were obtained. Among the top 20 genes, 13 common hub genes were identified. Further, area under the receiver operating characteristic curve (ROC AUC) analysis was performed to assess the genes as potential diagnostic blood biomarkers. Six genes, ESM1, FCN2, MDK, GPC3, CTHRC1 and COL6A6, exhibited an AUC value higher than 0.85 and were predicted as blood-secretory. This study highlights the potential of an integrative computational approach for discovering non-invasive blood-based biomarkers in LC, facilitating for further validation and clinical translation. SignificanceLiver cancer is one of the leading causes of premature death worldwide, with its prevalence and mortality rates projected to increase. Although current diagnostic methods are highly sensitive, they are invasive and unsuitable for repeated testing. Blood biomarkers offer a promising non-invasive alternative, but their wide dynamic range of protein concentration poses experimental challenges. Therefore, utilizing available omics data to develop a diagnostic model could provide a potential solution for accurate diagnosis. This study developed a computational method integrating machine learning and bioinformatics analysis to identify potential blood biomarkers. As a result, ESM1, FCN2, MDK, GPC3, CTHRC1 and COL6A6 biomarkers were identified, holding significant promise for improving diagnosis and understanding of liver cancer. The integrated method can be applied to other cancers, offering a possible solution for early detection and improved patient outcomes.

Derivative learning of tensorial quantities-Predicting finite temperature infrared spectra from first principles.

We develop a strategy that integrates machine learning and first-principles calculations to achieve technically accurate predictions of infrared spectra. In particular, the methodology allows one to predict infrared spectra for complex systems at finite temperatures. The method's effectiveness is demonstrated in challenging scenarios, such as the analysis of water and the organic-inorganic halide perovskite MAPbI3, where our results consistently align with experimental data. A distinctive feature of the methodology is the incorporation of derivative learning, which proves indispensable for obtaining accurate polarization data in bulk materials and facilitates the training of a machine learning surrogate model of the polarization adapted to rotational and translational symmetries. We achieve polarization prediction accuracies of about 1% for the water dimer by training only on the predicted Born effective charges.

Exploring the Accuracy of Machine Learning and Deep Learning in Engine Knock Detection

This study explores the use of machine learning for real-time detection of engine knocking, aiming to enhance early vehicle fault recognition. We extracted frequency modulation amplitude demodulation (FMAD) features from engine sound data and evaluated various machine-learning algorithms using MATLAB. The coarse decision tree algorithm emerged as the most effective, achieving a classification accuracy of 66.01%. Subsequently, by using deep learning models, we significantly improved the accuracy: a convolutional neural network (CNN) achieved 45.16%. accuracy, a deep learning recurrent neural network (RNN) model in LSTM achieved 90% accuracy, and further refinements pushed the accuracy to 93.55%. Additionally, we introduced a knock index to quantify noise levels during each engine cycle. This index, calculated from the integral of the absolute value of the first derivative of a band-pass-filtered vibration signal, provides a visual representation of knock strength. This approach shows promise for early detection of engine knocking, although further refinement of feature extraction methods and algorithm optimization is necessary for practical application. The study highlights the potential of integrating machine learning into real-time vehicle fault detection systems to improve their reliability and effectiveness.

Evaluating the efficacy of machine learning in calibrating low-cost sensors

Ambient air quality monitoring requires low-cost environmental sensor devices that are affordable and feasible for large-scale implementation. However, issues such as sensor drift, environmental sensitivity, and inter-sensor variability affect data accuracy and cannot be adequately addressed by traditional calibration methods. This paper summarizes the use of machine learning techniques for calibrating low-cost sensors. The literature review shows that machine learning models like Random Forest, Support Vector Regression, and Neural Networks significantly improve sensor accuracy and reliability. For instance, Random Forest models reduced the root mean squared error by 30% for PM2.5 measurements, while Neural Networks achieved an R value of 0.997 for methane sensors. Integrating machine learning with IoT and mobile technologies enhances real-time monitoring and spatial resolution. Identified gaps include the quality of training datasets, managing environmental variability, and improving model transferability across different contexts. Addressing these gaps through advanced models and real-time calibration methodologies will further enhance sensor performance, ensuring more precise and reliable environmental data.

ỨNG DỤNG PHƯƠNG PHÁP HỌC MÁY ĐỂ DỰ ĐOÁN CÁC YẾU TỐ MÔI TRƯỜNG TRONG THIẾT KẾ HỆ THỐNG THỦY CANH IOT

In IoT hydroponics, the integration of Internet of Things (IoT) and Machine Learning (ML) has opened up great opportunities to increase the productivity and smart system’s management. With the collected data from sensors in the environment, the machine learning model would analyze and predict the trend of the environmental factors, this combination not only improves the outcome quality but also helps saving the resources. Integrating machine learning into an IoT hydroponics system could not only create a smart, autonomous and adaptable system to changing conditions of environment in real-time but also optimize resources for a cost-effective and productive hydroponic system. In this study, a novel method was presented for predicting environmental factors using Machine Learning algorithm for smart IoT hydroponic systems. By applying this method, an IoT hydroponic system can predict the trends of environmental factors which affects the plants such as temperature, moisture, pH levels…. The experiment results show that the accuracy of the predicted data is reliable, it reached 94.2% for a day and 92.6% for a week. These results could help users take proactive measures to improve the cultivation quality.

An Integrative Machine Learning Approach in Detecting Gene Networks Predicting Liver Cancer Incidence and Survival

Introduction. Liver cancer (LC) is one of the most common types of cancer and the third largest cause of cancer-related deaths globally. Although there is no known cause of LC, the disease has some associated risk factors. Studies have also linked genetic mutations to LC. Patients who receive early diagnosis have a five-year relative survival rate of ~36%, whereas ~3% for late diagnoses.Methods. In this study, we employed a novel machine learning technique called netLDA which initially identified predictive strong DE genes and used as them hubs genes to identify coregulating genes that may only exhibit marginally weak DE effects. The detected genes form predictive gene networks (PGNs), which are used to make predictions.Results. Our study detected previously identified LC occurrence susceptibility strong genes; GPC3, STMN1 and TCF19. Potential biomarker genes for LC occurrence such as MT1E and PLVAP were detected as co-regulating weak genes. In our one-year survival study, SPP1 and ADH4 were detected as strong genes whilst MFAP4 was identified as novel co-regulating weak gene. The PGNs achieved an AUC of about 1.00 in our case-control study and 0.77 in our one-year survival study. Some of the detected genes overlapped with some significant LC related pathways to provide additional layers of biological significance.Conclusions. With our study, the identified predictive genes can propel studies into understanding the underlying mechanisms for LC development and prognosis. These identified biomarker genes can be used as new therapeutic targets by industries in the development of new drugs for LC.

Analysis of Instruments for Implementing Intelligent Job Matching Models

Vocational high schools in Indonesia face challenges in producing graduates who meet the needs of the job market. The link and match policy along with the Job Matching system are proposed to enhance curriculum relevance to the workforce. The intelligent job matching model, which integrates machine learning technology, is expected to provide job recommendations aligned with graduates' competencies. Data analysis techniques used include qualitative and quantitative statistical analysis, encompassing Validity Analysis, Practicality Analysis, and Effectiveness Analysis. The intelligent job matching model has proven effective in assisting in matching vocational high school students' competencies with workforce demands in the industry. In the 2023 job fair, the success ratio of students invited for interviews until the employment process reached 53.13%, exceeding the Ministry of education and culture's target of 40%. This indicates that the model can provide job recommendations accurately enough, aiding in improving the direct employment rate of vocational high school students. Research indicates that the intelligent job matching model is effective in facilitating students' transition to the workforce. The model has been tested for validity, practicality, and effectiveness in enhancing students' employability and strengthening the relationship between education and industry.

Enhancing healthcare facility resilience: utilizing machine learning model for airborne disease infection prediction

During this pandemic, advanced epidemiological models have been widely used to determine intervention strategies for controlling the spread of the disease in public and healthcare settings. These models played a crucial role by providing predictive insights into disease transmission dynamics, informing resource allocation and guiding policy decisions. However, the accuracy of these predictions depends on the substantial amount of input data, which was not readily available at the onset of the pandemic. Another concern with the existing models is their inability to adequately account for the complex indoor built environments, which has been shown to significantly impact infection risk. To tackle these issues, this paper discusses the potential of developing a joint modelling technique that integrates machine learning models, building information models and agent-based models to assess the risk of nosocomial airborne infections. With limited available data, machine learning models can determine infection risk with high confidence.

Machine Learning-Enhanced Benders Decomposition Approach for the Multi-Stage Stochastic Transmission Expansion Planning Problem

The necessary decarbonization efforts in energy sectors entail integrating flexible assets and increased levels of uncertainty for the planning and operation of power systems. To cope with this in a cost-effective manner, transmission expansion planning (TEP) models need to incorporate progressively more details to represent potential long-term system developments and the operation of power grids with intermittent renewable generation. However, the increased modeling complexities of TEP exercises can easily lead to computationally intractable optimization problems. Currently, most techniques that address computational intractability alter the original problem, thus neglecting critical modeling aspects or affecting the structure of the optimal solution. In this paper, we propose an alternative approach to significantly alleviate the computational burden of large-scale TEP problems. Our approach integrates machine learning (ML) with the well-established Benders decomposition to manage the problem size while preserving solution quality. The proposed ML-enhanced Multicut Benders Decomposition algorithm improves computational efficiency by identifying effective and ineffective optimality cuts via supervised learning techniques. We illustrate the benefits of the proposed methodology by solving multi-stage TEP problems of different sizes based on the IEEE24 and IEEE118 test systems, while also considering energy storage investment options..

Leveraging Topological Data Analysis and AI for Advanced Manufacturing: Integrating Machine Learning and Automation for Predictive Maintenance and Process Optimization

Leveraging Topological Data Analysis and AI for Advanced Manufacturing: Integrating Machine Learning and Automation for Predictive Maintenance and Process Optimization

Integrated machine learning models for predictive analysis of thermal and electrical power generation of a photo-thermal system at Catania, Italy

The Photovoltaic-Thermal (PV/T) system is designed for producing both electrical and thermal energy. Its efficiency is improved when temperature-sensitive PV cells are cooled using nanofluid coolants. However, the current models for predicting PV/T system performance with nanofluid cooling are limited. In order to fill this gap, this study aims to develop machine learning models to predict the electrical and thermal efficiencies of a water-based PV/T system. Three types of machine learning algorithms from the supervised learning method were selected in this study because of their ability to handle complex data and provide accurate predictions: the Multi-layer Perceptron (MLP), the Gradient Boosting Regressor (GBR) and the Light Gradient-Boosting Machine (LightGBM). Initially, the models are trained and validated using data consisting of 15,540 samples from a water-based PV/T system. On the basis of the results obtained, the MLP algorithm proved to be the best for the prediction of electrical energy, with an R2 value = 0.9906, against 0.9709 and 0.99 respectively for the other GBR and LGB algorithms, while the LGB algorithm proved effective for the prediction of thermal energy, with an R2 value = 0.983, against 0.983, 0.97 and 0.988 respectively for the GBR and LGB algorithms. Secondly, the best MLP and LGB models previously identified for the water-based PV/T system were adapted to predict, this time, the energy performance of an Ag/water nanofluid-based PV/T system. The results obtained show that these models are highly effective in predicting the electrical and thermal performance of the Ag/water nanofluid PV/T system, even in the absence of real data, making them very useful for real applications.

Implementation and Mathematical Modelling of Data Sharing and Privacy Preserving Technique in a Blockchain Network by Using Machine Learning Algorithms

Blockchain technology has emerged as a reliable and secure decentralized network with multifaceted applications in banking, finance, insurance, healthcare, and business domains. Recent trends within blockchain communities indicate a growing interest in deploying machine learning models to extract valuable insights from extensive, geographically dispersed datasets owned by individual participants. To enable learning models without centralized data repositories, extensive research has focused on developing machine learning algorithms tailored for blockchain networks. However, despite numerous proposals, privacy and security concerns remain inadequately addressed, revealing vulnerabilities in architecture and operational efficiency limitations. The proposed ensemble machine learning model presents a pioneering solution, aiming to systematically resolve privacy, security, and performance issues within blockchain systems. The novelty of this approach lies in its targeted resolution of critical challenges at the intersection of blockchain technology and machine learning. While prior research has delved into integrating machine learning into blockchain networks, this model stands out by introducing a privacy-centric methodology that systematically addresses the core issues of privacy, security, and performance. Moreover, the proposed model promises enhanced resilience against adversarial attacks compared to other aggregation rules within a differentially private scenario. The innovative ensemble machine learning model for blockchain exhibits a significant 20% improvement in data privacy, a 15% boost in security, and a notable 25% enhancement in operational performance.

Platelet Metabolites as Candidate Biomarkers in Sepsis Diagnosis and Management Using the Proposed Explainable Artificial Intelligence Approach.

Background: Sepsis is characterized by an atypical immune response to infection and is a dangerous health problem leading to significant mortality. Current diagnostic methods exhibit insufficient sensitivity and specificity and require the discovery of precise biomarkers for the early diagnosis and treatment of sepsis. Platelets, known for their hemostatic abilities, also play an important role in immunological responses. This study aims to develop a model integrating machine learning and explainable artificial intelligence (XAI) to identify novel platelet metabolomics markers of sepsis. Methods: A total of 39 participants, 25 diagnosed with sepsis and 14 control subjects, were included in the study. The profiles of platelet metabolites were analyzed using quantitative 1H-nuclear magnetic resonance (NMR) technology. Data were processed using the synthetic minority oversampling method (SMOTE)-Tomek to address the issue of class imbalance. In addition, missing data were filled using a technique based on random forests. Three machine learning models, namely extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and kernel tree boosting (KTBoost), were used for sepsis prediction. The models were validated using cross-validation. Clinical annotations of the optimal sepsis prediction model were analyzed using SHapley Additive exPlanations (SHAP), an XAI technique. Results: The results showed that the KTBoost model (0.900 accuracy and 0.943 AUC) achieved better performance than the other models in sepsis diagnosis. SHAP results revealed that metabolites such as carnitine, glutamate, and myo-inositol are important biomarkers in sepsis prediction and intuitively explained the prediction decisions of the model. Conclusion: Platelet metabolites identified by the KTBoost model and XAI have significant potential for the early diagnosis and monitoring of sepsis and improving patient outcomes.