Max Depth Research Articles

Implementation of The Extreme Gradient Boosting Algorithm with Hyperparameter Tuning in Celiac Disease Classification

Celiac Disease (CeD) is an autoimmune disorder triggered by gluten consumption and involves the immune system and HLA in the intestine. The global incidence ranges from 0.5%-1%, with only 30% correctly diagnosed. Diagnosis remains challenging, requiring complex tests like blood tests, small bowel biopsy, and elimination of gluten from the diet. Therefore, a faster and more efficient alternative is needed. Extreme Gradient Boosting (XGBoost), an ensemble machine learning technique that utilizes decision trees to aid in the classification of Celiac disease, was used. The aim of this study was to classify patients into six classes, namely potential, atypical, silent, typical, latent and none disease, based on attributes such as blood test results, clinical symptoms and medical history. This research method employs 5-fold cross-validation to optimize parameters that are max depth, n estimator, gamma, and learning rate. Experiments were conducted 96 times to get the best combination of parameters. The results of this research are highlighted by an improvement of 0.45% above the accuracy value with the default XGBoost parameter of 98.19%. The best model was obtained in the trial with parameters max depth of 3, n estimator of 100, gamma of 0, and learning rate of 0.3 and 0.5 after modifying the parameters, yielding an accuracy rate of 98.64%, a sensitivity rate of 98.43%, and a specificity rate of 99.72%. This research shows that tuning the XGBoost parameters for Celiac

Survivals of Titanic Prediction Utilizing Tree-based Machine Learning Models

The shipwreck of Titanic is a well-known tragedy. Although it happened more than a century ago, researchers are still investigating the patterns of the survivors to gain more insight into human behaviors in catastrophes. This paper adopts machine learning techniques, including decision tree, random forest, and gradient boosting, to conduct a binary classification to predict whether a person survived. The selected models are all tree-based, making it convenient to examine the importance of features. In the preprocessing stage, all numerical features are discretized. This paper first investigates the performances of the models. Subsequently, the model with the best performance generates and studies the importance of the feature. The result demonstrates that the decision tree classifier with a max depth equal to seven achieves the highest accuracy of 0.78. The results of the three models are similar, indicating that the research is robust. The feature importance generated by the decision tree classifier shows that sex and social status significantly impact the survival result. In addition, whether the person is a child also makes a difference. The discretized features do not have enough influence on the result of survival. This paper concludes that the tunned decision tree classifier is the best model to study the features in this paper, but the created features are not effective enough.

Phosphorus inactivation mitigates the effect of warm winters in a temperate shallow lake (Mielenko Lake, Poland)

Direct and indirect anthropopressure on water ecosystems is the serious problem throughout the world.. In the Northern Hemisphere, an increase in average air temperatures is observed, which implies the occurrence of a shorter period of snow and ice cover during the winter season. The winter 2019/2020 was unusual, because that was the first time in the record, that a complete lack of permanent ice cover was observed on numerous lakes in Poland. Such unusual conditions could influence lake functioning. Hence we analyzed the chemistry of the water–sediment interface (near-bottom and interstitial water and sediment) in the shallow, eutrophic Mielenko Lake (area 7.9 ha, max depth 1.9 m) in 2013 and 2019–2022 period to assess the influence of prolonged water circulation on the bottom zone. Mielenko Lake was subjected to a phosphorus inactivation procedure using Al and Fe salts (PAX 18, PIX 111) in 2020 and 2021. Our research revealed that unusually prolonged winter circulation caused a significant decrease in organic matter content in bottom sediment in 2020, as well as a decrease in NaOH-nrP fraction and TP amounts. That effect was short-term and it did not significantly influence the NaOH-rP fraction amounts. The released P was probably built in macrophytes biomass during vegetation season, because P inactivation has been limiting phytoplankton proliferation, and it favored shifting to a clearwater state with macrophytes domination. This was confirmed by decreasing in phytoplankton biomass, and a massive expansion of the macrophytes range noted in the second year of restoration. Our study shows, that P inactivation could mitigate the negative effects of warm winters in shallow lakes.

Machine Learning for Mapping and Forecasting Poverty in North Sumatera: A Data-Driven Approach

Discussing poverty is crucial because it affects many facets of society, including socioeconomic disparity, crime, and the inability to obtain high-quality education. One of the provinces with the highest poverty rate in Indonesia is North Sumatra. A strategy is required to gather accurate data to effectively reduce poverty. Poverty mapping and prediction were conducted in North Sumatra to get a precise spatial distribution of poverty, the operation of the poverty model, and forecasting using machine learning (ML). Poverty prediction was conducted using a random forest (RF) algorithm and poverty mapping was conducted using the K-Means algorithm. The poverty mapping showed a significant inertia value decline in the third and fourth clusters of the elbow graph. The third cluster (0.313) was superior to the fourth cluster (0.244) in the silhouette index. Thus, there were three poverty clusters - low, medium, and high - that were used in the model. The best model was created using the grid search cross-validation, while the best prediction results were created using the RF algorithm, with the following parameters: n-estimator = 50, max depth = 10, min samples split = 2, and min samples leaf = 1. The mean squared error (MSE) of the RF model's predictions was 0.002617, or satisfactory precision.

Predictive modeling of spring-back in pre-punched sheet roll forming using machine learning

This study outlines an experimental and computational endeavor aimed at developing a machine learning model to estimate spring-back values utilizing the decision tree methodology. A design of experiment approach was employed to collect a dataset, and based on the experimental results, a precise model was constructed to predict spring-back values. The model considered parameters such as thickness, diameter of circle hole, distance between the center hole and flange edge, and hole spacing. Various hyper parameters, including max depth and minimum samples for split, were explored, with configurations such as (30,5), (20,8), and (10,2) being evaluated to identify the optimal model for spring-back prediction. Analysis of the results demonstrated that the decision tree models accurately estimated spring-back values in cold roll forming of pre-punched sheets based on the input parameters. The coefficient of determination in the test section for decision tree models with parameters (30,5), (20,8), and (10,2) was found to be 0.90, 0.98, and 0.96, respectively. Additionally, the percentage of absolute error in the test section for the same decision tree models was calculated as 8.84%, 6.18%, and 7.6%, respectively.

Sedimentological characterization of earthquake-generated turbidites in fault-proximal glacial lakes: a case study from Jenny Lake, Teton range, Wyoming

Lakes in seismically active regions preserve valuable sedimentary archives of paleoseismic activity within their catchment and beyond. A series of glacially-excavated lakes positioned directly along the surface trace of the Teton normal fault at the base of the Teton Range, WY, are ideally situated to record past fault activity since their formation approximately 15,000 years ago. Here, we focus on the sediment fill contained in Jenny Lake (5 km2; approximately 73 m max depth) located at the bottom of Cascade Canyon, in the central Tetons, where postglacial slip rates are greatest. Past earthquakes that generated slope failures in and around Jenny Lake are expressed stratigraphically as coseismic turbidite deposits. These deposits were previously identified and dated in sub-bottom profiles and in sediment cores taken from multiple locations around the basin. In this study, we focus on the six thickest turbidites (ranging from 6 to 34 cm thick) present in multiple cores recovered from the central depositional basin and analyze them at sub-centimeter resolution for changes in physical, biological, and geochemical parameters, including sediment density, magnetic susceptibility, grain size distributions, organic content, and elemental composition. Results reveal each deposit contains a well-defined, three-component sedimentary sequence composed of a relatively homogenous and thick basal sandy unit with a sharp bottom contact, a similarly thick silt-rich middle unit, and a thin top unit of very fine-grained sediments. The characteristics of these components and their similarity between individual deposits suggest consistent sediment sources and transport pathways during successive earthquake events. Based on our analyses and the unique local geomorphic setting, we create a mechanistic model of coseismic turbidite formation in Jenny Lake, which may provide an improved framework for identifying, characterizing, and correlating earthquake-generated disturbance deposits in other Teton lakes and lakes in similar tectonic-geomorphological settings, for example, in the Basin and Range Province, USA or beyond.

Distribution of Normative Keratoconus (KC) indices for patients aged 6 to 21 Years using Pentacam HR in the Riyadh area

Objectives: To determine the distribution of Normative Keratoconus (KC) indices in normal Saudi patients aged between 6-21 years in the Riyadh area. Method: The present study employed a prospective, cross-sectional, observational, multicenter design to gather data from patients aged 6 to 21 who were seen in several non-ophthalmic emergency departments in Riyadh, Kingdom of Saudi Arabia, in 2017. Scheimpflug corneal tomography system in rotation was used to measure the cornea. Following optometric exams, corneal imaging and KC index measurements were performed using the Pentacam. Results: 366 study participants were analyzed after the exclusion criteria were applied. All children's mean and standard deviation (SD) ages were 17.41±4.83 (6–21). All the KC indices percentiles for pediatric, adolescent, and young adult groups are within normal ranges. Comparison between KC indices and age groups reveals statistically significant differences for most indices with P <0.05. As the age increases, the ISV, IVA, KI, CKI, IHA, IHD, BFS Front, BFS Back, ART Min, ART Max, ART Avg, and AC Depth decrease. Also, As the age increases, the topographic astigmatism (Topo Astig), PPI Min, PPI Max, and PPI Avg increase. There was a statistically significant correlation between KC indices and age, with most indices with P <0.05. The correlation coefficient ranges from 0.14 negative weak to 0.65 negative moderated, and from positive weak 0.22 to negative strong 0.82. Moreover, compression between KC indices and Topo Astig groups reveals statistically significant differences for most indices P <0.05. As the Topo astig increase, the K1, RSagMin PPIMax, and PPIAvg decrease. As the Topo Astig increases, the K2, ISV, IVA, CKI, IHA, BFS Back, and ACD increase. Also, there was a statistically significant correlation between Topo Astig groups and indices with P <0.05. The correlation coefficient ranges from 0.13 positive weak to 0.70 positive strong and from negative weak 0.11 to negative moderate 0.42. Conclusion: The KC indices of the current study were within normal range and comparable to prior studies. However, some indices were related to age, refractive error, and location. The normative values of multiple Pentacam indices showed statistically significant variations throughout the age groups and Topo astig subgroups.

Using Decision Tree Regression and Related Companies’ Stock Data to Predict Microsoft Stock Returns

Using AI machine learning (ML) models to predict stocks is a topic that has already been studied by the stock market research community. However, two ML methods have not been analyzed in the current literature regarding stock prediction: the decision tree regression analysis and the related company training data approach. Thus, this study will utilize both of these unfamiliar stock prediction methods to predict Microsoft’s stock returns. To begin with, the company data of Microsoft's biggest partners and competitors were imported from YahooFinance; this data was then used to form all the features for the stock prediction model (mean, standard deviation, price gaps, etc.). Next, the machine learning model was created using Python's Decision Tree Regressor method; the model was trained using data before 10/1/2001 and tested using data after 10/1/2001. Through repeatedly testing this model, hyperparameter tuning was performed to determine the model's best features and max depth for predicting Microsoft’s stock returns. In the end, the final prediction model reached a percentage accuracy (percentage of times correctly predicting stock return's direction) of 56.68%, and the plot (net returns using model vs. historical net returns) showed that model use resulted in more consistent and significantly higher net Microsoft stock returns. Therefore, this study demonstrated that both the Decision Tree Regressor and the related company training data approach are successful machine learning methods in predicting Microsoft's stock returns. However, further research is required to extend this study's results to other companies and/or different stock metrics.

Max Depth Impact on Heart Disease Classification: Decision Tree and Random Forest

Max Depth Impact on Heart Disease Classification: Decision Tree and Random Forest

Using Machine Learning to Predict Near-Term Thrombosis Risk in Patients with Polycythemia Vera

Introduction: Thrombosis (arterial and venous) remains the leading cause of morbidity/mortality for patients (pts) with polycythemia vera (PV) [1-2]. With appropriate therapy, thrombosis risk can be reduced to 1-5% per year (yr) [3]. Yet some seemingly well-controlled pts still suffer thrombotic events. This rate means that a well-powered clinical trial testing anti-thrombotic strategies in PV to improve thrombosis-free survival is prohibited by the large number of pts and study duration required. For example, the high-end annual incidence of thrombosis in ELN high-risk pts is <5%/yr, so a 2yr study hoping to demonstrate a 50% thrombosis reduction (α=5%, 1-ß=80%) requires 1104 pts. This study would be vastly more efficient if we could enroll only pts with high near-term thrombotic risk but we cannot accurately identify these pts. To address this, we developed a machine learning (ML) approach to dynamically predict this risk. Methods: After obtaining institutional approval, 526 pts with manually verified PV diagnoses were identified by querying electronic health records and our research database repository. The dataset (4M data elements,14K hematology clinic visits) includes lab measures, molecular/pathology studies, medications, and pt history. Data completeness varied with >90% availability in 200+ features. To assess the error introduced by imputation, we computed the Euclidean distance/Jaccard index between masked pre-imputed data and imputed data over 8K iterations. We found that features with <50% missingness could be imputed without dataset corruption. ML classification models (random forest (RF), boosted models, and support vector machines) were evaluated using AutoGluon. RF emerged as the top performer. Feature engineering, guided by Gini feature importance and clinical feedback, identified 21 training features. A parallelized pipeline optimized 1.6M model combinations (# of trees, node splitting features, max depth, thresholding) via grid search. Several optimization strategies were assessed. Of 470 pts with clinic visits through Mar 31, 2020, 75% were used for model training (Training Cohort) and 25% for model evaluation (Testing Cohort). Cohort details were previously reported [4]. A prospective Validation Cohort included clinic visits after Apr 1, 2020 and was unseen by trained models. A risk calculator designed in RShiny applied models corresponding to user-entered clinical data and computed 1yr thrombosis risk. User-entered data are not stored to protect data privacy. Results: Training/Testing pts had median age of 54, and 64 (14%) had a prior thrombosis at dx (8% venous, 7% arterial). Over a median followup of 10 yrs, 159 thromboses occurred in 115 pts (88 venous, 71 arterial). Annual incidence rates (IR) of thrombosis were higher shortly after dx (IR: 4% vs 1%, 2yr cutoff) and following a thrombotic event (IR: 10% vs 2%, 2yr cutoff). The full model performed very well (F1=0.91, AUC=0.84) when compared with ELN (F1=0.1, AUC=0.39) (Figure 1). High-performing simplified models (>90% testing/validation sensitivity) were applied to the Validation Cohort. The full 21-feature model and high-performing models (Figure 2) showed clear separation for 1yr thrombosis risk prediction. Features of importance included blood counts, body mass index (BMI), and time since dx were frequently represented in these models. Discussion: Over 1M models were trained on high-volume clinical data to predict near-term thrombosis risk in PV pts, considering intrinsic factors (age, blood type), disease events (time since dx/thrombosis), and short-term changes (BMI) to tailor risk mitigation strategies. Prospective validation demonstrates model generalizability. External validation using REVEAL study [1] and the Danish Registry data is ongoing. Our tool's ability to identify high-risk pt populations can be applied as an inclusion criterion for prospective clinical trials, reducing the time and pt numbers needed for a well-powered study. Whereas using ELN-defined high risk would require 1K pts to demonstrate a 50% reduction in thrombosis rates, our prospective ML model-defined high near-term risk would only require ~80 pts, making such a study very feasible.

Accident severity prediction modeling for road safety using random forest algorithm: an analysis of Indian highways.

Background: Road accidents claim around 1.35 million lives annually, with countries like India facing a significant impact. In 2019, India reported 449,002 road accidents, causing 151,113 deaths and 451,361 injuries. Accident severity modeling helps understand contributing factors and develop preventive strategies. AI models, such as random forest, offer adaptability and higher predictive accuracy compared to traditional statistical models. This study aims to develop a predictive model for traffic accident severity on Indian highways using the random forest algorithm. Methods: A multi-step methodology was employed, involving data collection and preparation, feature selection, training a random forest model, tuning parameters, and evaluating the model using accuracy and F1 score. Data sources included MoRTH and NHAI. Results: The classification model had hyperparameters 'max depth': 10, 'max features': 'sqrt', and 'n estimators': 100. The model achieved an overall accuracy of 67% and a weighted average F1-score of 0.64 on the training set, with a macro average F1-score of 0.53. Using grid search, a random forest Classifier was fitted with optimal parameters, resulting in 41.47% accuracy on test data. Conclusions: The random forest classifier model predicted traffic accident severity with 67% accuracy on the training set and 41.47% on the test set, suggesting possible bias or imbalance in the dataset. No clear patterns were found between the day of the week and accident occurrence or severity. Performance can be improved by addressing dataset imbalance and refining model hyperparameters. The model often underestimated accident severity, highlighting the influence of external factors. Adopting a sophisticated data recording system in line with MoRTH and IRC guidelines and integrating machine learning techniques can enhance road safety modeling, decision-making, and accident prevention efforts.

Predicting PM2.5 Concentrations Across USA Using Machine Learning

AbstractEconomic growth, air pollution, and forest fires in some states in the United States have increased the concentration of particulate matter with a diameter less than or equal to 2.5 μm (PM2.5). Although previous studies have tried to observe PM2.5 both spatially and temporally using aerosol remote sensing and geostatistical estimation, they were limited in accuracy by coarse resolution. In this paper, the performance of machine learning models on predicting PM2.5 is assessed with linear regression (LR), decision tree (DT), gradient boosting regression (GBR), AdaBoost regression (ABR), XGBoost (XGB), k‐nearest neighbors (K‐NN), long short‐term memory (LSTM), random forest (RF), and support vector machine (SVM) using PM2.5 station data from 2017 to 2021. To compare the accuracy of all the nine machine learning models, the coefficient of determination (R2), root mean square error (RMSE), Nash‐Sutcliffe efficiency (NSE), root mean square error ratio (RSR), and percent bias (PBIAS) were evaluated. Among all nine models, the RF (100 decision trees with a max depth of 20) and support vector regression (SVR; nonlinear kernel, degree 3 polynomial) models were the best for predicting PM2.5 concentrations. Additionally, comparison of the PM2.5 performance metrics displayed that the models had better predictive behavior in the western United States than that in the eastern United States.

Machine learning prediction of major adverse cardiac events after elective bariatric surgery.

Machine learning (ML) is an emerging technology with the potential to predict and improve clinical outcomes including adverse events, based on complex pattern recognition. Major adverse cardiac events (MACE) after bariatric surgery have an incidence of 0.1% but carry significant morbidity and mortality. Prior studies have investigated these events using traditional statistical methods, however, studies reporting ML for MACE prediction in bariatric surgery remain limited. As such, the objective of this study was to evaluate and compare MACE prediction models in bariatric surgery using traditional statistical methods and ML. Cross-sectional study of the MBSAQIP database, from 2015 to 2019. A binary-outcome MACE prediction model was generated using three different modeling methods: (1) main-effects-only logistic regression, (2) neural network with a single hidden layer, and (3) XGBoost model with a max depth of 3. The same set of predictor variables and random split of the total data (50/50) were used to train and validate each model. Overall performance was compared based on the area under the receiver operating curve (AUC). A total of 755,506 patients were included, of which 0.1% experienced MACE. Of the total sample, 79.6% were female, 47.8% had hypertension, 26.2% had diabetes, 23.7% had hyperlipidemia, 8.4% used tobacco within 1year, 1.9% had previous percutaneous cardiac intervention, 1.2% had a history of myocardial infarction, 1.1% had previous cardiac surgery, and 0.6% had renal insufficiency. The AUC for the three different MACE prediction models was: 0.790 for logistic regression, 0.798 for neural network and 0.787 for XGBoost. While the AUC implies similar discriminant function, the risk prediction histogram for the neural network shifted in a smoother fashion. The ML models developed achieved good discriminant function in predicting MACE. ML can help clinicians with patient selection and identify individuals who may be at elevated risk for MACE after bariatric surgery.

Hyperbaric treatment deviations for U.S. Navy divers: Spinal DCS

Introduction The United States Navy (USN) developed and refined standardized oxygen treatment tables for diving injuries, but USN tables may not address all situations of spinal decompression sickness (DCS). We describe a detailed recompression treatment regimen that deviated from standard USN protocol for an active-duty USN diver with a severe, delayed presentation of spinal cord DCS. Case Report A USN diver surfaced from his second of three dives on a standard Navy ‘no-Decompression’ Air SCUBA dive (Max depth 101 fsw utilizing a Navy Dive Computer) and developed mid-thoracic back pain, intense nausea, paresthesias of bilateral feet, and penile erection. Either not recognizing the constellation of symptoms as DCS and after resolution of the aforementioned symptoms, he completed the third planned dive (essentially an in-water recompression). Several hours later, he developed paresthesias and numbness of bilateral feet and legs and bowel incontinence. He presented for hyperbaric treatment twenty hours after surfacing from the final dive and was diagnosed with severe spinal DCS. Based on the severity of clinical presentation and delay to treatment, the initial and follow-on treatments were modified from standard USN protocol. MRI of the spine four days after initial presentation demonstrated a 2∙2 cm lesion at the T4 vertebral level extending caudally. Follow-up examinations over two years demonstrated almost complete return of motor and sensory function; however, the patient continued to suffer fecal incontinence and demonstrated an abnormal post-void residual urinary volume. An atypical presenting symptom, a discussion of MRI findings, and clinical correlations to the syndrome of spinal DCS are discussed throughout treatment and long-term recovery of the patient.

Variability in billfish vertical distribution and fishing interactions driven by environmental conditions in the Eastern Tropical Pacific Ocean

Abstract Blue marlin (Makaira nigricans) and sailfish (Istiophorus platypterus) are ecologically important predators and valuable species throughout the world’s recreational, commercial, and subsistence fisheries. Comparing multi-species vertical habitat use can inform ecological uncertainties such as inter-species competition, as well as relative vulnerabilities to fishing activities. In this study, we identified key differences in both depth use and which environmental variables drive these selections, which highlights the variability in the catchability both as target species in recreational fisheries and bycatch in commercial fisheries. To understand these two species’ vertical habitat use, we examined depth profiles from 26 sailfish and 48 blue marlin tagged with pop-up satellite archival tags deployed in the Eastern Tropical Pacific Ocean. While both species are surface-oriented, we found evidence of vertical niche partitioning where sailfish spend more time at deeper depths than blue marlin. Blue marlin recorded an average mean depth of 18.5 m (±10.8 m) during daytime and 5.2 m (±5.5 m) at nighttime (Figure 31a), while sailfish recorded an average mean depth of 23.6 m (±11.1 m) during daytime and 6.45 m (±4.64 m) at nighttime. Generalized additive mixed models fitted to predict mean and max depth revealed sea level anomaly (SLA), oxygen, sea surface temperature, and mixed layer depth as significant predictors of vertical habitat use for both species. We also examined catch logs from three recreational fishing lodges in Central America to understand the influence of environmental conditions on billfish sightings per unit effort. For blue marlin and sailfish, SLA was a significant predictor in each of the four depth models (mean day, mean night, max day, max night). SLA was the variable with highest percent deviance explained for all four sailfish depth models and three of the four blue marlin depth models and had a positive relationship with all response variables for all four blue marlin depth models and three sailfish depth models (mean daytime, max daytime, max nighttime), where higher positive SLA values were associated with deeper depth responses.

Parametric Analysis & Design Structure of Hydraulic Seed System

This paper is bases on the design and manufacturing of the sowing machine based on hydraulic principle named as ‘Hydraulic Seed Sowing Machine’. The machine is very efficient in working and will be able to easily sow the seeds like (bajra, rice, Val, gram, groundnut, pea, maize, groundnut, etc.). The machine is following the conditions such as – It operates on a permanent raised bed of 100 cm width which can sow seeds with fertilizer in 5 rows at 25 cm approximately apart. Seeds - bajra, rice, Val, gram, groundnut, pea, maize, groundnut, etc. Mechanism is preferably impact dibber and it make very thin slit of max depth of 30-40 cm. It covers the seeds after sowing and be operated by 1-2 people either 1 hp power motor or manually and sow 0.6 to 1 ha in approximately 6 hours. Hydraulic Seed Sowing Machine is Potential of being self-driven to get higher speed

Accident severity prediction modeling for road safety using random forest algorithm: an analysis of Indian highways

Background: Road accidents claim around 1.35 million lives annually, with countries like India facing a significant impact. In 2019, India reported 449,002 road accidents, causing 151,113 deaths and 451,361 injuries. Accident severity modeling helps understand contributing factors and develop preventive strategies. AI models, such as random forest, offer adaptability and higher predictive accuracy compared to traditional statistical models. This study aims to develop a predictive model for traffic accident severity on Indian highways using the random forest algorithm. Methods: A multi-step methodology was employed, involving data collection and preparation, feature selection, training a random forest model, tuning parameters, and evaluating the model using accuracy and F1 score. Data sources included MoRTH and NHAI. Results: The classification model had hyperparameters ’max depth’: 10, ’max features’: ’sqrt’, and ’n estimators’: 100. The model achieved an overall accuracy of 67% and a weighted average F1-score of 0.64 on the training set, with a macro average F1-score of 0.53. Using grid search, a random forest Classifier was fitted with optimal parameters, resulting in 41.47% accuracy on test data. Conclusions: The random forest classifier model predicted traffic accident severity with 67% accuracy on the training set and 41.47% on the test set, suggesting possible bias or imbalance in the dataset. No clear patterns were found between the day of the week and accident occurrence or severity. Performance can be improved by addressing dataset imbalance and refining model hyperparameters. The model often underestimated accident severity, highlighting the influence of external factors. Adopting a sophisticated data recording system in line with MoRTH and IRC guidelines and integrating machine learning techniques can enhance road safety modeling, decision-making, and accident prevention efforts.

Possibilities of Improving Water Quality of Degraded Lake Affected by Nutrient Overloading from Agricultural Sources by the Multi-Point Aeration Technique

This research concerns the restoration of the strongly eutrophied Lake Łajskie (Masurian Lake District, Poland) that received pollutants from the agricultural catchment. It is a shallow (max depth 7.4 m) and small (area 48 ha) reservoir with a wide profundal zone characterized by complete deoxidation in summer. Due to its important natural and social role, the lake has undergone a restoration process. Artificial oxygenation is the main method of improving water quality. Due to unfavorable morphometric conditions, the necessity to use multi-point aeration was assumed. An experimental installation based on eight reactors selectively oxygenating only the over-bottom waters was launched in 2019. In 2021, spatial studies of the effectiveness of aerators’ work were carried out at 96 measuring points grouped into 12-test fields around each reactor. The investigations were performed three times during the summer season. It was shown that, in the water of the deepest layers of Lake Łajskie, the oxygen concentration around all reactors increased to an average level of 1–2 mg L−1. The oxygenation efficiency was varied and related to the distance from the aerator and the location in relation to the direction of oxygenated water outflow. The results of the research indicate the legitimacy of designing multi-point oxygenation systems in the restoration of waterbodies degraded as a result of the impact of agriculturally transformed catchments.

Prediction of Software Defects using Ensemble Machine Learning Techniques

During software development and maintenance, predicting software bugs becomes critical. Defect prediction early in the software development life cycle is an important aspect of the quality assurance process that has received a lot of attention in the previous two decades. Early detection of defective modules in software development can support the development team in efficiently and effectively utilizing available resources to provide high-quality software products in a short amount of time. The machine learning approach, which works by detecting hidden patterns among software features, is an excellent way to identify problematic modules. The software flaws in NASA datasets MC1, MW1, KC3, and PC4 are predicted using multiple machine learning classification algorithms in this work. A new model was developed based on altering the parameters of the previous XGBoost model, including N_estimator, learning rate, max depth, and subsample. The results were compared to those obtained by state-of-the-art models, and our model outperformed them across all datasets.

A Weekday and Weekend Separation-Based Algorithm for Short-Term Load Forecasting

This paper presents a methodology for forecasting hourly loads over seven days, using a separation of weekday and weekend data. The data used for the forecasting included the date, weather, and load data from January 1, 2021, with the forecast being conducted for four weeks from November 1 to 28, 2021. The methodology involved separating the data into weekdays and weekends, conducting separate forecasts for each, selecting features for use in each process, and optimizing the forecasting model's parameters through hyperparameter tuning. To select features, the Shapley Additional Explanations method was used to calculate feature importance, and Pearson correlation coefficients were used to measure linearity between features. This allowed for the selection of input features with high importance while avoiding those with high linearity with other features. The parameters of the forecasting model were optimized through Grid search, and the optimal combination of the XGBoost forecasting model's Learning Rate, n estimators, and Max depth was assessed. After all processes were completed, forecasting was performed, and the forecasting error was presented through Normalized Mean Absolute Error, Mean Absolute Percentage Error, and Normalized Root Mean Squared Error. The results obtained using the proposed algorithm were compared with those obtained without using it. In future studies, the authors plan to investigate forecasting for special dates, such as public holidays, and explore the use of forecasting algorithms that consider the probability of input data.