Reliable Data Research Articles

Improving Air Quality Data Reliability through Bi-Directional Univariate Imputation with the Random Forest Algorithm

Forecasting the future levels of air pollution provides valuable information that holds importance for the general public, vulnerable populations, and policymakers. High-quality data are essential for precise and reliable forecasts and investigations of air pollution. Missing observations arise when the sensors utilized for assessing air quality parameters experience malfunctions, which result in erroneous measurements or gaps in the dataset and hinder the data quality. This research paper presents a novel approach for imputing missing values in air quality data in a univariate approach. The algorithm employs the random forest (RF) algorithm to impute missing observations in a bi-directional (forward and reverse in time) manner for air quality (particulate matter less than 2.5 μm (PM2.5)) data from the Republic of Serbia. The algorithm was evaluated against simple methods, such as the mean and median imputation methods, for missing observations over durations of 24, 48, and 72 h. The results indicate that our algorithm yielded comparable error rates to the median imputation method for all periods when imputing the PM2.5 data. Ultimately, the algorithm’s higher computational complexity proved itself as not justified considering the minimal error decrease it achieved compared with the simpler methods. However, for future improvement, additional research is needed, such as utilizing low-code machine learning libraries and time-series forecasting techniques.

Penerapan Teknologi Sidik Jari dalam Sistem Hukum: Tinjauan terhadap Prinsip – Prinsip Pembuktian

The application of fingerprint technology in the legal system has brought significant changes to the evidentiary process. This technology enables the identification of individuals with high accuracy, which is an important element in law enforcement and dispute resolution. This study explores how fingerprint technology is applied in the Indonesian legal context, reviews relevant evidentiary principles, and analyzes the implications of its use in various legal cases. This research is descriptive analytical, meaning that this research aims to describe certain circumstances or situations of the objects or events being studied without intending to draw conclusions that can be applied generally. The approach used in this research is the normative juridical method. The results show that, although fingerprint technology makes a significant contribution to the effectiveness of evidentiary systems, fingerprint technology has long been used in the legal system as a tool for individual identification. Fingerprints, with their unique characteristics in each person, provide a guarantee of highly accurate identification. This technology is not only used in crime but also in various aspects of government and public services, such as population registration and securing access. The application of fingerprint technology in a modern legal context is becoming increasingly relevant with advances in digital technology that enable fast and efficient data processing. There are challenges related to privacy, data reliability and the integrity of legal processes. This research also proposes recommendations to increase the integration of fingerprint technology in the legal system in accordance with the principles of justice and human rights.

Pengaruh Motivasi dan Disiplin Kerja Terhadap Kinerja Karyawan CV. Santoso Teknik Karawang

The main objective of this research is to assess the individual and combined impact of work discipline and work motivation on employee performance at CV Santoso Teknik Karawang. Descriptive and verification strategies were used in the quantitative analysis of this research. Fifty people from CV Santoso Teknik Karawang became the population or research subjects. Sampling using saturated sampling means conducting a survey of all workers at CV. Santoso Teknik Karawang has fifty employees. Data were collected via a Google Forms survey in which various items were rated using a Likert scale. Various data validity and reliability tests, as well as classic assumption tests are available in the SPSS version 22 application. Multiple linear regression, t-test, f-test, and coefficient of determination are part of this series of tests. Motivation and work discipline partially influence employee performance, based on the results of the t test. Apart from that, the f test confirms that there is a significant simultaneous influence between motivation and work discipline on the performance of CV Santoso Teknik Karawang..

A dataset of NDVI for the vegetation growing season in Central Asia with a resolution of 250 m (2001–2020) and 30 m (2020)

Central Asia is the largest arid and semi-arid region in the Northern Hemisphere, and its ecological environment is extremely fragile and susceptible to the effects of global climate change. Maintaining the stability of the region's ecosystem is crucial to global economic and social development due to its unique geographic location. Vegetation serves as a significant indicator of the ecological environment, and the spatial and temporal distribution pattern, as well as the change trend of vegetation, are important indicators in evaluating the ecological status of the region. The Normalized Difference Vegetation Index (NDVI) is a commonly used remote sensing index to study vegetation, which characterizes the spatio-temporal changes of vegetation. This dataset utilized MODIS13Q1 to generate long-term time series growing season mean NDVI data with a spatial resolution of 250m in Central Asia from 2001 to 2020. To obtain the 2020 growing season mean NDVI data with a higher spatial resolution of 30m, the Cubist algorithm based on rule segmentation regression was utilized to fuse the Landsat data and MODIS data. Meanwhile, this dataset carried out quality control of data products from three aspects: data source quality control, model training optimization, and model independent verification to ensure the accuracy and reliability of data. The generation of this dataset provides powerful data support for the analysis of vegetation dynamic change and spatial pattern in Central Asia.

Sensor Fault Detection and Classification Using Multi-Step-Ahead Prediction with an Long Short-Term Memoery (LSTM) Autoencoder

The Internet of Things (IoT) is witnessing a surge in sensor-equipped devices. The data generated by these IoT devices serve as a critical foundation for informed decision-making, real-time insights, and innovative solutions across various applications in everyday life. However, data reliability is often compromised due to the vulnerability of sensors to faults arising from harsh operational conditions that can adversely affect the subsequent operations that depend on the collected data. Hence, the identification of anomalies within sensor-derived data holds significant importance in the IoT context. This article proposes a sensor fault detection method using a Long Short-Term Memory autoencoder (LSTM-AE). The AE, trained on normal sensor data, predicts a 20-step window, generating three statistical features via SHapley Additive exPlanations from the estimated steps. These features aid in determining potential faults in the predicted steps using a machine learning classifier. A secondary classifier identifies the type of fault in the sensor signal. Experimentation on two sensor datasets showcases the method’s functionality, achieving fault detection accuracies of approximately 93% and 97%. It is possible to attain a perfect fault classification performance by slightly modifying the feature calculation approach. In a univariate prediction scenario, our proposed approach demonstrates good fault detection and classification performance.

Abstract P406: Transforming Clinical Practices for Blood Pressure Management in Community Health Centers through Timely Data and Technical Support

Introduction: Blood pressure (BP) control can prevent major adverse cardiac events and is monitored through various means, including the Centers for Medicare and Medicaid Services 165 metric for BP control - an annual reporting metric for community health centers (CHCs) within the Uniform Data System framework that becomes accessible 6 months after the conclusion of the reporting period. In pursuit of more immediate indicators related to BP management, the National Hypertension Control Initiative (NHCI) initiated collaboration with a population health platform to support select CHCs with rapidly available data on 19 BP-related metrics. Hypothesis: With technical support and regular access to data, CHCs can implement timely clinical practice and process changes. Methods: To support CHCs in using the population health platform, NHCI staff provided technical assistance in a rapid cycle performance improvement (RCPI) to 10 interested CHCs. This opportunity allowed CHCs to partner with NHCI staff to review BP related metrics in the population health platform and to identify opportunities for clinical process and practice changes. Over an 8-week period from October to December 2023, 7 CHCs were actively involved in a series of 5-6 structured 30–60-minute meetings with both NHCI registered nurse and quality improvement staff. These meetings reviewed BP metrics, mapped clinical workflows for improvement, and supported data analytics. CHCs often chose 1-2 BP metrics for focus, concentrating on incremental change and short-term improvement. Results: Among participating CHCs that reported, they revealed successful or anticipated adjustments in team responsibilities and clinical processes, including leadership involvement and new methods. Limitations included CHC staff engagement, time, concerns with data reliability, and familiarity with the population health platform. Most reporting CHCs shared that their goals were greatly impacted by access to readily available performance data and that they plan to continue using the population health platform to affect performance moving forward. Conclusions: For participating CHCs, individualized guidance based on timely data may have been helpful in analyzing workflows and procedures, assessing clinical care practices, and implementing quality improvement plans. Frequent data may allow CHCs to make clinical practice and process changes to efficiently measure impact and more quickly improve BP control than standard methods.

Accurate pulse time distribution determination using MLEM algorithm in integral experiments

Integral experiments play a crucial role in advancing nuclear science and technology by providing critical data that validate theoretical models and enhance reactor designs. This study presents a novel approach to accurately determine pulse time distribution in integral experiments conducted with pulsed accelerators. By strategically placed monitors and shields at angles of 0° and 90° relative to the beam direction, neutron flight times from the target are measured, and a response matrix for neutron emission at different times is constructed through simulation. The Maximum Likelihood Expectation Maximization (MLEM) algorithm is employed for pulse time reconstruction, with the gamma ray flight time spectrum from monitors used as the initial spectrum to streamline the computational process. Experimental validation using a standard polyethylene sample and n-p scattering cross-sections confirms the accuracy of the method. Results are compared across multiple nuclear databases such as CENDL-3.2, ENDF/B-VIII.0, JENDL-5.0, and JEFF-3.3 libraries. The developed method significantly enhances the precision of pulse time distribution determination, thereby improving the quality and reliability of experimental data obtained from integral experiments conducted with pulsed accelerators.

Spatiotemporal variations of permafrost extent in Mongolia during 1950–2022

Permafrost in Mongolia is located in the transition zone between high-latitude and high-altitude permafrost regions of the Northern Hemisphere, with large temperature differences and complex subsurface characteristics. In this study, the reliability of the skin temperature data from the ERA5-Land product covering Mongolia is assessed via site observations. The ERA5-Land skin temperature dataset shows a cold bias, which is more pronounced in the cold season. Following calibration based on elevation differences, significant improvements are observed at both the annual scale (92 % improvement in RMSE (root mean square error) and 98 % improvement in MBE (mean bias error)) and the seasonal scale (78 % improvement in RMSE and 82 % improvement in MBE). Additionally, the spatial variations in the surface freezing index (SFI) and surface thawing index (STI) are most pronounced in the central and northeastern Mongolia. The SFI exhibits a significant decreasing trend of 7.16 °C·d/year, while the STI shows a significant increasing trend of 4.49 °C·d/year. Furthermore, the permafrost extent in Mongolia is simulated from 1950 to 2022 using the frost number (Fn) model and the temperature on top of permafrost (TTOP) model. The validated results indicate that the accuracy of the Fn model is relatively high, with an overall accuracy of 0.9 and a Kappa coefficient of 0.47. The permafrost extent in Mongolia has declined from 734.7 × 103 km2 in the 1950 s to 480.1 × 103 km2 in the 2010 s, with a prominent decrease of 3.2 × 103 km2/decade after 1994. According to the variations in permafrost extent during past 72 years, the Hovsgol and Khentii Mountain ranges have experienced significant permafrost degradation.

The ancient mitochondrial genome of a Siberian roe deer (Capreolus pygargus) from Jartai Pass Site in Xinjiang, China, and its phylogenetic relationships

The ancient mitochondrial genome of a Siberian roe deer (Capreolus pygargus) coded as NJ26S from Jartai Pass Site was obtained by high throughput sequencing. The damage pattern demonstrated the authenticity and reliability of the ancient DNA data. The length of the mitogenome was 16,357 bp, which contained 13 protein-coding genes, two rRNA genes, 22 tRNA genes, and one control region. The total base composition of the mitochondrial genome is 28.17% A, 25.01% T, 11.89% G, 19.72% C, and 15.21% missing data with an AT composition of 53.18%. A maximum-likelihood phylogenetic tree was recovered including other roe deer sequences under the TIM2 + I + G4 model. This study presents molecular evidence indicating the presence of Capreolus pygargus in the Xinjiang Uygur Autonomous Region in China more than 3,000 years ago.

Bitcoin Price Prediction Using Deep Bayesian LSTM With Uncertainty Quantification: A Monte Carlo Dropout–Based Approach

ABSTRACTBitcoin, being one of the most triumphant cryptocurrencies, is gaining increasing popularity online and is being used in a variety of transactions. Recently, research on Bitcoin price predictions is receiving more attention, and researchers have investigated the various state‐of‐the‐art machine learning (ML) and deep learning (DL) models to predict Bitcoin price. However, despite these models providing promising predictions, they consistently exhibit uncertainty, which cannot be adequately quantified by classical ML models alone. Motivated by the enormous success of applying Bayesian approaches in several disciplines of ML and DL, this study aims to use Bayesian methods alongside Long Short‐Term Memory (LSTM) to predict the closing Bitcoin price and consequently measure the uncertainty of the prediction model. Specifically, we adopted the Monte Carlo dropout (MC‐Dropout) method with the Bayesian LSTM model to quantify the epistemic uncertainty of the model's predictions and provided confidence intervals for the predicted outputs. Experimental results showed that the proposed model is efficient and outperforms other state‐of‐the‐art models in terms of root mean square error (RMSE), mean absolute error (MAE) and R2. Thus, we believe that these models may assist the investors and traders in making critical decisions based on short‐term predictions of Bitcoin price. This study illustrates the potential benefits of utilizing Bayesian DL approaches in time series analysis to improve data prediction accuracy and reliability.

Enhancing 5G energy efficiency through digital twin networks: A comprehensive review

AbstractIn the era of 5G, seamless mobility handovers are critical in densely populated regions like Malaysia to mitigate disruptions and inefficiencies. 5G networks offer unprecedented data speeds and reliability, essential for advancing mobile communication and Internet of Things applications. However, ensuring continuous connectivity and service during mobility remains challenging, especially in urban settings. Digital twin technology presents a promising solution to enhance 5G handover mechanisms. A digital twin network (DTN) mirroring Malaysia's 5G infrastructure is proposed, utilising real‐time data and user behaviour insights to optimise energy consumption during handovers. The focus is on energy‐efficient protocols and algorithms, reviewed through a systematic literature review. The DTN aims to enhance mobility handover efficiency through predictive handovers and adaptive resource allocation, bolstered by sustainable practices such as edge computing. The potential of DTNs to optimise 5G handover processes is explored, starting with the foundational concepts of 5G mobility and digital twins, highlighting the need for improved strategies in high‐mobility scenarios. Methodologies leveraging digital twins to predict network conditions, simulate handover scenarios, and proactively manage decisions are examined, reducing latency and packet loss. Case studies demonstrate how digital twins adapt dynamically to network changes and user mobility, thereby improving quality of service and user experience. Malaysia's specific 5G mobility challenges are addressed with a tailored DTN emphasising energy efficiency, evaluated through practical applications. Evaluation criteria assess effectiveness with in‐depth analysis of methods, performance metrics, limitations, and recommendations for future research. Challenges and future directions including scalability, security, and real‐time data processing, are discussed, aiming to integrate digital twin technology with 5G networks for enhanced connectivity. This abstract provides a roadmap for leveraging digital twins to optimise 5G network performance sustainably, guiding future research and implementation strategies.

Exploring Effective Therapeutic Approaches for COVID-19: A Review on Progress and Prospects

: Global health and economy have been severely impacted by the COVID-19 pandemic that was brought on by the SARS-CoV-2 coronavirus. This makes the creation of potent medications for the treatment of COVID-19 disease a top goal. In clinical trials, several medications that have been repurposed for the treatment of COVID-19 have shown promise. Nevertheless, there are a lot of obstacles to overcome in the creation of COVID-19 pharmacological therapy. The inability to pinpoint prospective treatment targets is one difficulty caused by the unclear etiology of COVID-19. Another difficulty is the virus' quick evolution, which can result in the creation of drug-resistant variants. Furthermore, the quick start of clinical trials has been prompted by the strong demand for effective therapies. A lack of reliable data on the safety and efficacy of medications can result from the early start of clinical trials that have been prompted by the strong demand for effective medicines. Despite these obstacles, the development of various promising pharmacological treatments for COVID-19 has advanced. These include immune-modulating medications, like dexamethasone and tocilizumab, as well as antiviral medicines, like remdesivir and favipiravir. Combination therapies using several medications may also be useful in enhancing outcomes for COVID-19 patients. Despite some encouraging advancements, there are still a lot of obstacles to be addressed in the development of pharmacological therapy for COVID-19. Further research is needed to identify the most effective treatment approaches for this disease.

A pioneering approach to measure increased resilience to face climate change: insights from the Race to Resilience campaign

This paper illustrates a methodology to measure the impact of resilience-building actions on the increased resilience of people and natural systems to face climate change, developed and field-tested around the Race to Resilience Campaign. Despite increasing acknowledgment of the need for robust methodologies and indicators to monitor and evaluate efforts across adaptation planning and implementation, and provide credibility, accountability and transparency to such actions, there is still a lack of sufficiently standardized and agreed upon metrics able to capture the effect of resilience-building actions. The proposal illustrated in this manuscript offers a pioneering approach for high-level tracking, monitoring and evaluation of resilience-building efforts of non-state actors, based on two complementing sets of metrics: depth metrics measure the degree to which an action is generating a change to fundamental conditions which can demonstrably be related to increasing resilience; while magnitude metrics offer a quantification of the beneficiaries that are affected by these changes. Underlying both stand the Resilience Attributes: properties which can be soundly associated with triggering resilience across different systems, and which can then be used to assess increased resilience ‘by proxy’: that is, by seeing how an action sets forth changes in properties commonly associated with resilience. These Attributes were identified based on updated scientific literature and co-construction exercises with global experts. The integration of Depth and Magnitude indices, adjusted by a Confidence Index evaluating data reliability, allows to estimate the overall contribution of a set of actions on increasing resilience against climate challenges. Based on the above, a possible Monitoring & Evaluation cycle is proposed, and an illustration is offered on two case studies from the Race to Resilience campaign. Key strengths, lessons learned and insights are summarized to stimulate the global discussion, in the context of the Global Stocktake and Global Goal on Adaptation.

Enhancing Supply Chain Resilience: the Impact of Blockchain & Emerging Technologies on Traceability

Supply chain traceability is an invaluable tool for companies to gain a competitive advantage today. The ever-changing global business environment, unpredictable market conditions, and changing consumer preferences call for agile supply chain systems. The pandemic helped reveal the vulnerabilities within the supply chain networks that need addressing. Various challenges, such as data reliability and integration, affect the current supply chain traceability and must be addressed. Such factors lead to inefficiencies within the system, which affect the process and lead to inconsistencies. The paper uses a systematic literature review to identify recent challenges and solutions to supply chain traceability. An analysis of blockchain technology's role in promoting transparency and security in supply chain systems was conducted. Lastly, an in-depth analysis of how improving traceability can help the supply chain networks build resilience.

Data quality control system and long-term performance monitor of LHAASO-KM2A

The KM2A is the largest sub-array of the Large High Altitude Air Shower Observatory (LHAASO). It consists of 5216 electromagnetic particle detectors (EDs) and 1188 muon detectors (MDs). The data recorded by the EDs and MDs are used to reconstruct primary information of cosmic-ray and gamma-ray showers. To ensure the reliability of the LHAASO-KM2A data, a three-level quality control system has been established. It is used to monitor the status of detector units, stability of reconstructed parameters and the performance of the array based on observations of the Crab Nebula and Moon shadow. This paper will introduce the control system and its application on the LHAASO-KM2A data collected from August 2021 to July 2023. During this period, the pointing and angular resolution of the array were stable. From the observations of the Moon shadow and Crab Nebula, the results achieved using the two methods are consistent with each other. For example, according to the observation of the Crab Nebula with KM2A at energies from 25 TeV to 100 TeV, the time averaged pointing errors are estimated to be −0.003°±0.005° and 0.001°±0.006° in the R.A. and Dec directions, respectively.

Estimating Adverse Events Associated With Herbal Medicines Using Pharmacovigilance Databases: Systematic Review and Meta-Analysis.

Herbal medicines (HMs) are extensively used by consumers/patients worldwide. However, their safety profiles are often poorly reported and characterized. Previous studies have documented adverse events (AEs) associated with HMs, such as hepatotoxicity, renal failure, and allergic reactions. However, the prevalence rate of AEs related to HMs has been reported to be low. To date, no systematic review and meta-analysis has comprehensively analyzed the AEs of HMs using published data acquired from pharmacovigilance (PV) databases. This study aimed to (1) estimate the reporting rate of the AEs of HMs using PV databases and (2) assess the detailed data provided in AE reports. In this systematic review and meta-analysis, MEDLINE/PubMed, SCOPUS, EMBASE, and CINAHL were systematically searched for relevant studies (until December 2023). The DerSimonian-Laird random effects model was used for pooling the data, and subgroup analyses, the meta-regression model, and sensitivity analysis were used to explore the source of heterogeneity. Crombie's checklist was used to evaluate the risk of bias (ROB) of the included studies. In total, 26 studies met the eligibility criteria. The reporting rate of the AEs of HMs ranged considerably, from 0.03% to 29.84%, with a median overall pooled estimate of 1.42% (IQR 1.12%-1.72%). Subgroup analyses combined with the meta-regression model revealed that the reporting rate of the AEs of HMs was associated with the source of the reporter (P=.01). None of the included studies provided full details of suspected herbal products, only the main ingredients were disclosed, and other potentially harmful components were not listed. This systematic review and meta-analysis highlighted risks related to HMs, with a wide range of reporting rates, depending on the source of the reporter. Continuous efforts are necessary to standardize consumer reporting systems in terms of the reporting form, education, and follow-up strategy to improve data quality assurance, aiming to enhance the reliability and utility of PV data for monitoring the safety of HMs. Achieving effective monitoring and reporting of these AEs necessitates collaborative efforts from diverse stakeholders, including patients/consumers, manufacturers, physicians, complementary practitioners, sellers/distributors, and health authorities. PROSPERO (Prospective International Register of Systematic Reviews) CRD42021276492.

Adversarial Variational Autoencoders to Extend and Improve Generative Model

Generative artificial intelligence (GenAI) has been advancing with many notable achievements like ChatGPT and Bard. The deep generative model (DGM) is a branch of GenAI, which is preeminent in generating raster data such as image and sound due to the strong role of deep neural networks (DNNs) in inference and recognition. The built-in inference mechanism of DNN, which simulates and aims at synaptic plasticity of the human neuron network, fosters the generation ability of DGM, which produces surprising results with the support of statistical flexibility. Two popular approaches in DGM are the variational autoencoder (VAE) and generative adversarial network (GAN). Both VAE and GAN have their own strong points although they share and imply the underlying theory of statistics as well as significant complex via hidden layers of DNN when DNN becomes effective encoding/decoding functions without concrete specifications. This research unifies VAE and GAN into a consistent and consolidated model called the adversarial variational autoencoder (AVA) in which the VAE and GAN complement each other; for instance, the VAE is a good data generator by encoding data via the excellent ideology of Kullback–Leibler divergence and the GAN is a significantly important method to assess the reliability of data as to whether it is real or fake. In other words, the AVA aims to improve the accuracy of generative models; besides, the AVA extends the function of simple generative models. In methodology, this research focuses on the combination of applied mathematical concepts and skillful techniques of computer programming in order to implement and solve complicated problems as simply as possible.

ACWA Power’s CSP Performance Assessment

A CSP Plant and in particular the solar field has hundreds of thousands of components and the corresponding process parameters. This needs to be monitored and evaluated every day to make sure that the plant is working efficiently. A performance assessment with a production simulation is done daily where the expected and actual power production are compared. Such evaluation is usually done with the plant performance model and the necessary data is compiled in calculation programs like Excel. The work is repetitive, time consuming and human errors may easily occur. Typically, the evaluations are done from different parties with deviating results and conclusions. ACWA Power has developed an online simulation platform to automatize the data processing to increase the data reliability and reduce the time effort. The application has been implemented in the Noor Energy 1 hybrid solar plant in Dubai and shall be presented in this paper.

A detailed methodology for estimating health-related hazards of workplace exposure to indoor formaldehyde vapours

A comprehensive risk assessment method was applied to examine the risks associated with airborne formaldehyde occupational exposure among hospital laboratory staff. The method assessed exposure levels and health impacts by integrating area and personal air sampling, biological monitoring, and self-reported health data. Samples were collected from 74 workplaces across various departments using NIOSH method 3500 and were analyzed via UV-vis spectrophotometry. The data showed significant differences in exposure levels between departments (p≤0.05) and confirmed the efficacy of the method in identifying risk differences. Despite average personal exposure levels being measured lower than occupational limits, individual assessments indicated that some participants surpassed these limits, emphasizing the necessity of personal monitoring for workers with higher risks. The high prevalence of respiratory symptoms, such as cough and wheezing among staff, indicated the need for further investigation and targeted interventions. Although estimated cancer and non-cancer risks were within safe thresholds, the study emphasized the importance of continuous exposure monitoring and the implementation of effective control measures in hospital laboratory departments with formaldehyde emission. This integrated method improved the reliability and generalizability of formaldehyde exposure risk assessments and aided in the development of safe occupational health practices.•The method integrated personal and area sampling with advanced calibration for precise occupational exposure evaluation in laboratories.•The method used of biomarkers to assess formaldehyde absorption in the body estimating both cancerous and non-cancerous health risks associated with occupational exposure.•Addressed traditional method limitations and integrated risk components to improve data reliability for workplace safety and health risk management.

Stream salinity prediction in data-scarce regions: Application of transfer learning and uncertainty quantification

Scarcity of stream salinity data poses a challenge to understanding salinity dynamics and its implications for water supply management in water-scarce salt-prone regions around the world. This paper introduces a framework for generating continuous daily stream salinity estimates using instance-based transfer learning (TL) and assessing the reliability of the synthetic salinity data through uncertainty quantification via prediction intervals (PIs). The framework was developed using two temporally distinct specific conductance (SC) datasets from the Upper Red River Basin (URRB) located in southwestern Oklahoma and Texas Panhandle, United States. The instance-based TL approach was implemented by calibrating Feedforward Neural Networks (FFNNs) on a source SC dataset of around 1200 instantaneous grab samples collected by United States Geological Survey (USGS) from 1959 to 1993. The trained FFNNs were subsequently tested on a target dataset (1998-present) of 220 instantaneous grab samples collected by the Oklahoma Water Resources Board (OWRB). The framework's generalizability was assessed in the data-rich Bird Creek watershed in Oklahoma by manipulating continuous SC data to simulate data-scarce conditions for training the models and using the complete Bird Creek dataset for model evaluation. The Lower Upper Bound Estimation (LUBE) method was used with FFNNs to estimate PIs for uncertainty quantification. Autoregressive SC prediction methods via FFNN were found to be reliable with Nash Sutcliffe Efficiency (NSE) values of 0.65 and 0.45 on in-sample and out-of-sample test data, respectively. The same modeling scenario resulted in an NSE of 0.54 for the Bird Creek data using a similar missing data ratio, whereas a higher ratio of observed data increased the accuracy (NSE = 0.84). The relatively narrow estimated PIs for the North Fork Red River in the URRB indicated satisfactory stream salinity predictions, showing an average width equivalent to 25 % of the observed range and a confidence level of 70 %.