Target Variables Research Articles

Multi-Step Forecasting of Chlorophyll Concentration with Multi-Attention Collaborative Network

In a marine environment, the concentration of chlorophyll is an important indicator of quality, which is also considered an indicator used to predict the marine ecological environment, which is further considered an important means of predicting red tide disasters. Although existing methods for predicting chlorophyll concentration have achieved encouraging performance, there are still two limitations: (i) they primarily focus on the correlation between variables while ignoring negative noise from non-predictive variables and (ii) they are unable to distinguish the impact of chlorophyll from that of non-predictive variables on chlorophyll concentration at future time points. In order to overcome these obstacles, we propose a Multi-Attention Collaborative Network (MACN)-based triangle-structured prediction system. In particular, the MACN consists of two branch networks, with one named NP-net, focusing on non-predictive variables, and the other named T-net, applied to the target variable. NP-net incorporates variable-distillation attention to eliminate the negative effects of irrelevant variables, and its outputs are used as auxiliary information for T-net. T-net works on the target variable, and both its encoder and decoder are related to NP-net to use the output of NP-net for assistance in learning and prediction. Two actual datasets are used in the experiments, which show that the MACN performs better than various kinds of state-of-the-art techniques.

A Selective Under-Sampling (SUS) Method for Imbalanced Regression

Many mainstream machine learning approaches, such as neural networks, are not well suited to work with imbalanced data. Yet, this problem is frequently present in many real-world data sets. Collection methods are imperfect, and often not able to capture enough data in a specific range of the target variable. Furthermore, in certain tasks data is inherently imbalanced with many more normal events than edge cases. This problem is well studied within the classification context. However, only several methods have been proposed to deal with regression tasks. In addition, the proposed methods often do not yield good performance with high-dimensional data, while imbalanced high-dimensional regression has scarcely been explored. In this paper we present a selective under-sampling (SUS) algorithm for dealing with imbalanced regression and its iterative version SUSiter. We assessed this method on 15 regression data sets from different imbalanced domains, 5 synthetic high-dimensional imbalanced data sets and 2 more complex imbalanced age estimation image data sets. Our results suggest that SUS and SUSiter typically outperform other state-of-the-art techniques like SMOGN, or random under-sampling, when used with neural networks as learners.

Subthreshold nanosecond laser therapy for chronic central serous chorioretinopathy: A prospective study.

To investigate the morphologic and functional outcomes of nanosecond subthreshold (ST) laser treatment for patients with chronic central serous chorioretinopathy (CSC). In this prospective study, 44 patients were treated with the ST nanosecond laser with a follow-up period of 12 months. All target variables were measured at 1, 3, 6 and 12 months after the first laser treatment. This study showed a significant improvement in macular sensitivity (MS), a significant reduction in central macular thickness (CMT) as well as a significant reduction in subretinal fluid height (SRF) after 3 months of treatment. The subfoveal choroidal thickness (SFCT) was significantly reduced after 12 months of treatment. However, the best-corrected visual acuity (BCVA) (logMAR) did not change significantly at any time during the study. A high proportion of patients (85%) showed complete resolution of SRF after 12 months, indicating a positive response to treatment in the majority of our patients. This study showed statistically significant functional improvement in MS as well as significant anatomical reduction in CMT, SFCT and SRF height in CSC patients treated with ST nanosecond laser therapy. Patients with higher SRF at baseline required repeated laser treatments to achieve complete resolution of the SRF.

What-if: A causal machine learning approach to control-oriented modelling for building thermal dynamics

Operational optimization of buildings can improve efficiency, and reduce costs and emissions. This optimization typically relies on a model of the building thermal dynamics, which is used by a predictive controller to obtain optimal schedules for heating, ventilation and air conditioning. This model has historically been constructed by domain experts using physical properties of the building. However, this approach scales poorly as more and more residential and commercial buildings need to be modelled and controlled. As a consequence, researchers and practitioners have turned to data-driven models, trained only on observational data. However, this alternative is no panacea: such models often fail to generalize to truly unseen conditions due to their inability to learn cause–effect relationships - i.e. they are not causal, rather they only learn correlational associations between input and target variables. In this paper, we demonstrate this problem using classical machine learning models trained on data from two different use cases (a simulated RC building and nine real-world Dutch buildings). Our results show that, unlike commonly used data-driven methods, causal machine learning (CML) algorithms trained on debiased data can produce accurate models necessary for control-oriented applications which outperform baseline models by over 40%, besides learning the correct causal associations which we verify using a custom testing environment as well as SHAP feature analysis. These results emphasize the need to move beyond simplistic data-driven methods if control-oriented applications are to be realized in a feasible manner.

Mobile Apps for Musculoskeletal Diseases: A Scoping Review

In recent decades, health apps have become widespread. Patients with musculoskeletal diseases use this software to self-manage their symptoms and continue their diet. The purpose of this study is to scope a review of articles related to the apps developed for musculoskeletal patients to identify and synthesize relevant knowledge. This research was based on the population, intervention, comparison, and outcome (PICO) model. A preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram was used to report the literature search and review process. The Cochrane Risk of Bias tool was used to evaluate bias in randomized clinical trials (RCTs). A database search was conducted to find RCTs published from January 1, 2012, to November 29, 2022, across databases. For each article relevant to an app for patients with musculoskeletal disease, we qualitatively evaluated its information on country, target health problem, outcomes, comparisons, study design, and variables assessed to determine the efficacy of the app. Among the reviewed articles, 34 were included in this review. In these articles, most apps investigated focused on knee osteoarthritis (32.35%). The apps addressed a variety of variables, but self-management in pain control (38.23%) and exercise (17.64%) were the top two subjects. Most of the apps were developed in the US (23.52%). The global trend toward mHealth apps for self-management of diseases is increasing. It seems that existing mobile apps not only are insufficient but also do not fully cover the growing needs of patients with musculoskeletal disorders. Continued research and development in this area are critical to meet the needs of a growing population living with the long-term effects of musculoskeletal diseases.

Ranking the determinants of financial performance using machine learning methods: an application to BIST energy companies

Energy has been a key driver of change globally. As a developing country, Türkiye's increasing energy demand and consumption highlight the growing importance of efficient and sustainable energy management for its future. This study aims to determine the variables of the financial performance of 12 energy companies. Three different models are created with the return on assets, return on equity, and net profit margin as financial performance indicators of 12 firms. 12 financial ratios are used as input variables as determinants of financial performance. In the analysis, 37 quarterly data between 2014Q4-2023Q4 are used as the sample period. In machine learning, 17 different algorithms are considered in the selection of the appropriate model. The findings indicate that the Bagged Tree algorithm achieved successful outcomes for the ROA target variable, the Boosted Tree model demonstrated effective performance for the ROE model, and the Linear SVM algorithm yielded favorable results for the NPM model. According to the result obtained by the LIME method, Liquidity Ratio and Cash Ratio affect the ROA, ROE, and NPM models positively, while inventory turnover affects the models negatively.

Assessing credit risks through a scoring model with time-varying parameters

Subject. This article discusses the method of dynamic modeling, namely the building of a scoring model with time-varying parameters. Objectives. The article aims to develop a scoring model with time-varying parameters for assessing credit risks using the data of a microfinance organization as a case study. Methods. For the study, I used a critical review of the literature on the use of dynamic modeling and a combination of classical logistic regression and time series models Results. Based on the analysis of data from a microfinance organization, the article confirms that changes in independent variables may contain a trend. The obtained confirmation is evidence of the temporal dependence of true parameters when using the developed model, which indicates the high efficiency of the scoring model in the framework of assessing the credit risks of a microfinance organization. Cyclical changes in the tested target and independent variables have been empirically proven, which makes it possible to conclude that the true parameters are time-dependent within the framework of the model used. Conclusions and Relevance. The empirical experiment shows that the developed method can significantly improve the efficiency of scoring models in the implementation of a credit risk analysis system in financial institutions. A scoring model with time-varying parameters can be used in the risk management system of any financial institutions.

Exploring correlates of high psychiatric inpatient utilization in Switzerland: a descriptive and machine learning analysis

Abstract Background This study investigated socio-demographic, psychiatric, and psychological characteristics of patients with high versus low utilization of psychiatric inpatient services. Our objective was to better understand the utilization pattern and to contribute to improving psychiatric care. Methods One-hundred and twenty inpatients of the University Psychiatric Clinics (UPK) Basel, Switzerland, participated in this cross-sectional study. All patients were interviewed using different clinical scales. As target variables we investigated the number of days of psychiatric inpatient treatment within a 30-month period. Results Despite including multiple relevant patient variables and using elaborate statistical models (classic univariate und multiple regression, LASSO regression, and non-linear random forest models), the selected variables explained only a small percentage of variance in the number of days of psychiatric inpatient treatment with cross-validated R2 values ranging from 0.16 to 0.22. The number of unmet needs of patients turned out to be a meaningful and hence potentially clinically relevant correlate of the number of days of psychiatric inpatient treatment in each of the applied statistical models. Conclusions High utilization behavior remains a complex phenomenon, which can only partly be explained by psychiatric, psychological, or social/demographic characteristics. Self-reported unmet patient needs seems to be a promising variable which may be targeted by further research in order to potentially reduce unnecessary hospitalizations or develop better tailored psychiatric treatments.

Monitoring of the trough concentration of valproic acid in pediatric epilepsy patients: a machine learning-based ensemble model.

Few personalized monitoring models for valproic acid (VPA) in pediatric epilepsy patients (PEPs) incorporate machine learning (ML) algorithms. This study aimed to develop an ensemble ML model for VPA monitoring to enhance clinical precision of VPA usage. A dataset comprising 366 VPA trough concentrations from 252 PEPs, along with 19 covariates and the target variable (VPA trough concentration), was refined by Spearman correlation and multicollinearity testing (366 × 11). The dataset was split into a training set (292) and testing set (74) at a ratio of 8:2. An ensemble model was formulated by Gradient Boosting Regression Trees (GBRT), Random Forest Regression (RFR), and Support Vector Regression (SVR), and assessed by SHapley Additive exPlanations (SHAP) analysis for covariate importance. The model was optimized for R2, relative accuracy, and absolute accuracy, and validated against two independent external datasets (32 in-hospital and 28 out-of-hospital dataset). Using the R2 weight ratio of GBRT, RFR and SVR optimized at 5:2:3, the ensemble model demonstrated superior performance in terms of relative accuracy (87.8%), absolute accuracy (78.4%), and R2 (0.50), while also exhibiting a lower Mean Absolute Error (9.87) and Root Mean Squared Error (12.24), as validated by the external datasets. Platelet count (PLT) and VPA daily dose were identified as pivotal covariates. The proposed ensemble model effectively monitors VPA trough concentrations in PEPs. By integrating covariates across various ML algorithms, it delivers results closely aligned with clinical practice, offering substantial clinical value for the guided use of VPA.

Role of life and usage characteristics, technical attributes and crowdsourced product reviews in smartphone exchange prices

Purpose Smartphone demand has been driving people towards refurbished electronic products. However, a lack of transparency in refurbished product pricing makes purchases time-consuming and reduces customer trust. Thus, our research aims to help practitioners and researchers understand how product life and usage characteristics, technical attributes and crowdsourced product reviews and sentiments affect exchange prices for refurbished/remanufactured smartphones.Design/methodology/approach Our five-stage exchange price predictive framework begins with data gathering and predictor variable identification. Thereafter, customer review data were scraped to populate both customer ratings and textual content, enabling sentiment analysis for the various smartphone configurations. Stepwise regression was used to find statistically significant factors and validate the predictive model. Testing for nonlinear effects, normality, outliers and homoskedasticity warrants power transformation of the target variable. The analysis used data from GSMArena.com and Amazon.com.Findings Our study validates extant findings and provides several novel insights for functional yet hedonistic products like smartphones. Unlike other pure hedonistic products, refurbished phone buyers care more about usage duration than life. Besides having a strong affinity for the sleekness of the phone, such customers are strongly dissuaded by the presence of negative textual content in the customer reviews.Originality/value Our study augments the current understanding of exchange price modelling by bringing in perspectives from life cycle characteristics, technical attributes and product reviews.

EVALUATING THE EFFECTIVENESS OF MACHINE LEARNING ALGORITHMS IN PREDICTING CRYPTOCURRENCY PRICES UNDER MARKET VOLATILITY: A STUDY BASED ON THE USA FINANCIAL MARKET

The cryptocurrency market is one of the most dynamic and volatile markets in the world's financial ecosystem, and investment landscapes in the US financial market have changed so much. In slightly over a decade, cryptocurrencies have moved from niche digital assets to mainstream investment opportunities such as Bitcoin, Ethereum, and many others. The prime objective of this research project was to investigate the effectiveness of various machine learning algorithms in the prediction of cryptocurrency prices within the volatile US financial market. This research pinpointed which Machine Learning techniques provide the most accurate and reliable predictions under different market conditions, with a full understanding of their strengths and limitations. The dataset gathered for analyzing and forecasting cryptocurrency prices entailed diverse and extensive data points, affirming a well-rounded foundation for machine learning algorithms. Particularly, current and historic price data from cryptocurrency exchanges such as Binance, Coinbase, and Kraken, together with trading metrics important for the definition of market dynamics. Aggregated data from financial databases such as Coin-Market-Cap, Crypto-Compare, and Yahoo Finance comes in structured form and presents historical consistency, hence perfectly fitting for machine learning applications. Models considered for the study ranged from simple, linear methods to complex ensemble and gradient-boosting algorithms. Precise performance evaluation is a proxy of its reliability and correctness of effectiveness in price predictions in a cryptocurrency market. Several measures of the effectiveness of prediction have been used here for assessing the different properties of models' performance: Precision, Recall, and F1-Score. Additional performance metrics were applied to evaluate the models in this study including Mean Absolute Error, Root Mean Squared Error, and R-squared. The gradient Boosting model did an excellent job as compared to other algorithms, as the values of accuracy, precision, recall, and F1-score for both classes were quite high. All three models have quite a relatively low MAE and RMSE, which means that each model is remarkably good at predicting the target variable. The application of machine learning models in the sphere of cryptocurrency price prediction might finally give very important implications to investors and stakeholders of the financial market in the USA, especially since recently, cryptocurrencies have been made integral parts of both individual and institutional investors' portfolios and trading strategies. To investors, it may provide indications of the entry and exit points, diversification of portfolios, and risk management by using machine learning models. Consolidation with the financial system will indeed mark a strategic shift toward data-driven decision-making in investment management and trading by integrating machine learning models into the financial systems.

Prediction of manhole-cover skid-resistance performance using decision trees and regression analysis

This study conducts skid-resistance coefficient tests on 3,600 manhole covers and collects data on various influencing factors, such as surface conditions, pattern types, starting zones, turning zones, usage duration, wheel track locations, and surrounding traffic volume. A preliminary classification analysis of the data is performed using decision trees to identify key factors affecting manhole-cover skid resistance. Subsequently, a multivariate nonlinear regression model is applied to examine the impact of these factors on the target variables. By integrating decision trees and nonlinear regression in a hybrid modeling approach, the study aims to reveal complex patterns in the data and generate reliable predictive results. The decision-tree analysis using non-continuous variables reveals that manhole covers located at turns and starting zones, and those with worn surfaces are more likely to fail to meet safety standards for skid resistance. The multivariate nonlinear regression model, which compares the predicted values against the actual British Pendulum Number (BPN) test results, achieves an accuracy of 82.9%.

Genomic predictors of drug sensitivity in cancer: Integrating genomic data for personalized medicine in the USA

Despite applying conventional predictive methodologies to obtain genomic insights, predicting drug sensitivity for healthcare organizations in the USA remains a daunting challenge. Cancer is a highly dynamic, adaptive disease tumor cells have repeatedly shown the capability to evolve mechanisms whereby therapeutic interventions can be evaded. Besides, one genomic alteration seldom predicts drug sensitivity. This research project aimed to address the challenges of predicting drug sensitivity by leveraging the GDSC dataset, an extensive resource connecting genomic profiles of cancer cell lines with their sensitivity to a wide range of anti-cancer drugs. This research's key focus was identifying robust genomic markers, including any specific mutations, gene expression patterns, or epigenetic modifications associated with drug sensitivity or resistance. Advanced machine learning and statistical methods were utilized by the predictive models to analyze complex relations that may exist between different genomic alterations and their drug sensitivity. The dataset used for this research project was derived from the Kaggle website. This dataset was compiled by the research project Genomics of Drug Sensitivity in Cancer collaboration between the Sanger Institute in the United Kingdom and the Massachusetts General Hospital Cancer Center in the United States. In their investigation, there was a massive screening of human cancer cell lines with a wide range of anti-cancer drugs. Data collection was performed by large-scale screening of diverse anti-cancer drugs against human cancer cell lines of various types. Cell viability was measured using the Cell-Titer-Glo assay following 72 hours of drug treatment. Several machine learning models were deployed, namely, Random Forest, Linear Regression, and XG-Boost, which exhibited specific strengths. Specific performance metrics used included MSE, RMSE, MAE, and R². As the statistics indicate, among the three models, Random Forest stands out and performs the best on this dataset across all metrics. A smaller value of MAE, MSE, and RMSE signifies that it provided the best forecast for the target variable. It also gave the highest R-squared value. Application of drug sensitivity prediction analysis in cancer can provide an overview of the mechanisms that underlie both tumor response and resistance by investigating the model predictions. The proposed predictive models have the potential to make significant impacts on clinical decision-making in cancer therapy. Predictive models derive informed decisions regarding a patient's risk of recurrence of disease, their response to certain therapies, and their prognosis based on complex clinical and genomic data. Keywords: Genomic Predictors; Drug Sensitivity; Genomic Markers; Personalized medicine; Machine Learning; Random Forest Algorithm.

Simulating flood risk in Tampa Bay using a machine learning driven approach

Machine learning (ML) models can simulate flood risk by identifying critical non-linear relationships between flood damage locations and flood risk factors (FRFs). To explore it, Tampa Bay, Florida, is selected as a test site. The study’s goal is to simulate flood risk and identify dominant FRFs using historical flood damage data as target variable, with 16 FRFs as predictor variables. Five different ML models such as decision tree (DT), support vector machine (SVM), adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), and random forest (RF) were adopted. RF classifies 2.42% of Tampa Bay as very high risk and 2.54% as high risk, while XGBoost classifies 3.85% as very high risk and 1.11% as high risk. Moreover, the communities reside at low altitudes and near the waterbodies, with dense man-made infrastructure, are at high flood risk. This study introduces a comprehensive framework for flood risk assessment and helps policymakers mitigate flood risk.

An explainable AI framework for spatiotemporal risk factor analysis in public health: a case study of cardiovascular mortality in South Korea

ABSTRACT Understanding environmental disease risk factor analysis at the district level is essential for gaining valuable insights into regional disease variations, offering a broader perspective compared to individual-level studies. Recently, explainable artificial intelligence (XAI) has received increasing attention in the analysis of factors affecting public health. However, previous purely data-driven XAI-based risk factor analyses faced challenges in capturing regional effect of environmental variables, leading to confusion regarding key spatiotemporal risk factors. Therefore, this study proposes a framework that includes two complementary XAI-based risk factor analyses following two assumptions. Regionally rescaled environmental variables must account for the unequal effects on environmental factors, which are likely influenced by variations in adaptation capacity to weather conditions and differences in exposure-response relationships to air pollutants. District-level disease distribution highlights geographic disparity in sociodemographic vulnerability, whereas temporal variation in diseases by district underscores temporal environmental impacts. Based on these two hypotheses, we rescaled environmental variables using two complementary schemes: one that employs the district-level disease distribution as the target variable, and another that utilizes the temporal residual of the disease within each district. We evaluated this framework by analyzing the association between cardiovascular age-standardized mortality rate (CVD-ASMR) and various risk factors in South Korea from 2010 to 2019, using high-performing random forest and light gradient boosting models with additive Shapley explanation. Compared to previous purely data-driven XAI-based analyses, the proposed schemes achieved significantly better results in capturing regional exposure-response relationships. In two complementary schemes, the most explainable factor to districts with high CVD-ASMR was low education level related to sociodemographic vulnerability, whereas the most explainable factors to high temporal CVD-ASMR patterns were low greenness and high air pollution levels. In addition, the two complementary schemes enabled us to reasonably analyze the interaction effect of the two risk factors, i.e. temperature and air pollutants. Furthermore, high CVD-ASMR and its high temporal variation were observed in situations of high sociodemographic vulnerability with poor air quality. These findings provide insightful public health planning for sustainable cities and society by pinpointing high-risk areas and tailoring strategies to address regional environmental challenges.

Bayesian hierarchical probabilistic forecasting of intraday electricity prices

We address the need for forecasting methodologies that handle large uncertainties in electricity prices for continuous intraday markets by incorporating parameter uncertainty and using a broad set of covariables. This study presents the first Bayesian forecasting of electricity prices traded on the German intraday market. Endogenous and exogenous covariables are handled via Orthogonal Matching Pursuit (OMP) and regularising priors. The target variable is the IDFull price index, with forecasts given as posterior predictive distributions. Validation uses the highly volatile 2022 electricity prices, which have seldom been studied. As a benchmark, we use all intraday transactions at the time of forecast to compute a live IDFull value. According to market efficiency, it should not be possible to improve on this last-price benchmark. However, we observe significant improvements in point measures and probability scores, including an average reduction of 5.9% in absolute errors and an average increase of 1.7% in accuracy when forecasting whether the IDFull exceeds the day-ahead price. Finally, we challenge the use of LASSO in electricity price forecasting, showing that OMP results in superior performance, specifically an average reduction of 22.7% in absolute error and 20.2% in the continuous ranked probability score.

ASSESSING THE CONTRIBUTION OF CONTAINMENT MEASURES, VACCINATION COVERAGE AND MOBILITY IN THE EVOLUTION OF DEATHS FROM COVID-19 IN BRAZIL

This work proposes a method to rank the contribution of containment measures, vaccination coverage and mobility to contain the evolution of the COVID-19 pandemic in different states of Brazil. The proposed method applies the automatic learning of regression models using decision trees through the XGBoost algorithm. To interpret it globally, the SHapley Additive exPlanations (SHAP) was used, which is an algorithm based on Shapley's cooperative game theory to quantify the contribution of the analyzed characteristics to the evolution of the target variable (deaths). The evaluation results of the method indicated its efficiency to quantify the contribution of each variable in a robust way. It reveals that the percentages of vaccine coverage of the first and second dose, in addition to the closure of schools, were the measures that had the greatest contribution to the evolution of the number of deaths from COVID-19. The weighting of the variables can help the responsible actors in the elaboration of public policies to minimize the socioeconomic effects in their regions. Since, in countries with great social inequality, the use of only a few more efficient measures would be less harmful than a lockdown, as it would be extremely harmful to the quality of life of these poorer populations.

Longitudinal Analysis of Risk Factors for Pulmonary Function Decline in Chronic Lung Diseases Over Five Years.

Chronic lung diseases (CLDs) are a major global health concern, characterized by a progressive decline in pulmonary function that severely impacts quality of life. It is essential to identify and predict the primary risk factors for CLDs. This study aims to establish a predictive model to assist healthcare providers in the early identification of high-risk patients and timely interventions and treatment options. This study utilized questionnaire data from the China Health and Retirement Longitudinal Study (CHARLS) collected in 2011, 2013, and 2015. A latent class growth model (LCGM) was established using CLDs as the baseline sample. This model stratified the patients based on the extent of the decline in Δpeak expiratory flow (ΔPEF), which served as the target variable. Independent variables included age, gender, smoking status, body mass index, education level, and comorbidities. A random forest model was developed using Python, and the importance of the feature was visualized through the SHAP method. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curve analysis, calibration curve analysis, and decision curve analysis. After screening, a total of 553 patients with CLDs were included in the study. The random forest model pinpointed grip strength, age, education level, gender, and asthma as the top five risk factors for pulmonary function decline. Specifically, the model demonstrated robust predictive performance with an area under the ROC curve (AUC) value of 0.77, affirming its accuracy and clinical applicability. Both calibration and decision curves further substantiated the reliability of the model in identifying patients at increased risk for pulmonary function decline. The predictive model developed in this study serves as a valuable tool for clinicians to target early interventions and optimize treatment strategies to enhance the quality of care and patient outcomes in the management of CLDs.

Enhancing Wind Power Forecasting Accuracy with Hybrid Deep Learning and Teaching-Learning-Based Optimization

Forecasting wind power generation is crucial for ensuring grid security and the competitiveness of the power market. This paper presents an innovative approach that combines deep learning (DL) with Teaching-Learning-Based Optimization (TLBO) to predict wind power output accurately. Using a real dataset spanning diverse weather conditions and turbine specifications collected between January 2018 and March 2020, the study employs 18 features as inputs, including Ambient Temperature, Wind Direction, and Wind Speed, with real power output in kW as the target variable. Metaheuristic algorithms including Particle Swarm Optimization (PSO), Barnacles Mating Optimizer (BMO), Biogeography-Based Optimization (BBO), and Firefly Algorithm (FA) are comprehensively compared for model optimization. TLBO-DL consistently provides forecasts that closely align with actual wind power values across instances, substantiated by its low RMSE of 98.7601, indicating effective minimization of errors in wind power forecasting. Comparative analysis with other algorithms reveals that TLBO-DL outperforms PSO-DL (RMSE: 102.6627), BMO-DL (RMSE: 132.4839), BBO-DL (RMSE: 103.8517), and FA-DL (RMSE: 104.7282) in terms of overall forecasting accuracy. The variations in the performance of other algorithms across instances highlight the robustness and effectiveness of TLBO-DL in achieving accurate wind power forecasts. Overall, TLBO-DL emerges as a reliable and superior algorithm for wind power forecasting, consistently providing accurate forecasts across a range of instances.

An economic evaluation of sensor-assisted health monitoring in dairy farming using the example of a rumen bolus.

The study investigates the economics of sensor-assisted dairy health management and indicates a certain economic potential in the use of a commercial rumen bolus capable of tracking activity and core body temperature. The economic evaluation was performed applying a stochastic model with the net return (NR) of investment of the sensor system as the target variable. The calculated NR considers the gross margin (GM) for both sensor-assisted and visual health monitoring, time savings through sensor-assisted monitoring, additional time spent addressing false positive messages from the sensor system, labor costs, and all costs associated with the investment in the sensor system. The analysis relies on a data set acquired from a dairy research and demonstration farm on which 65 dairy cows were equipped with the sensor system. A comparison of health-related messages issued by the rumen bolus with disease diagnoses shows that the sensor system issued a message in 7 of 11 cases of retained placenta (sensitivity = 64%), in 19 of 31 cases of clinical hypocalcemia (sensitivity = 61%), in 30 of 70 cases of mastitis (sensitivity = 43%), in 6 of 24 cases of metritis (sensitivity = 25%), and in 2 of 42 cases of diseases of the locomotor system (sensitivity = 5%) in a defined observation period, in many cases several days before the visual diagnosis. SimHerd (A/S Viborg, Denmark) was applied to determine the GM as a function of incidence, somatic cell count, risk of a mild case of disease, and days of milk withdrawal. In a workshop, veterinarians (n = 9) used the data set to assess the impact of using the sensor system on these parameters. The empirical distributions given by the veterinarians' individual assessments were used to model the parameters considered in the calculation of the "sensor-assisted" GM. For the modeled Holstein herds with a milk yield of 9,000 kg, simulation results show that average NR of investment ranges from +€23 to +€119 per cow per year for a herd of poor health, from -€12 to +€84 per cow per year for a herd of average health, and from -€33 to +€63 per cow per year for a herd of good health, depending on the scenario. The assumptions made regarding changes in labor had a strong influence on the calculated NR of investment. For a full economic evaluation of the sensor system, other functions (estrus detection, calving detection) and functional extensions (e.g., monitoring rumination) have to be considered.