Inaccurate Estimation Research Articles

Objective Bayesian trend filtering via adaptive piecewise polynomial regression

Several methods have been developed for nonparametric regression problems, including classical approaches such as kernels, local polynomials, smoothing splines, sieves, and wavelets, as well as relatively new methods such as lasso, generalized lasso, and trend filtering. This study proposes an objective Bayesian trend filtering method based on model selection. The procedure followed in this study estimates the functions based on adaptive piecewise polynomial regression models with two components. First, we determine the intervals with varying trends using Bayesian binary segmentation and then evaluate the most reasonable trend via Bayesian model selection at these intervals. This trend filtering procedure follows Bayesian model selection that uses intrinsic priors, which eliminated any subjective input. Additionally, we prove that the proposed method using these intrinsic priors was consistent when applied to large sample sizes. The behavior of the proposed Bayesian trend filtering procedure is compared with the trend filtering using a simulation study and real examples. Finally, we apply the proposed method to detect the variance change points under mean changes, whereas the existing methods yielded inaccurate estimates of the variance change points when the mean varied smoothly, as the sudden-change assumption was violated in such cases.

Response prediction of self-centered concrete walls using artificial neural networks

ABSTRACT In this research, responses of self-centered concrete walls were predicted by developing an efficient machine learning model, namely artificial neural network (ANN). A database based on 51 experimental data of self-centered walls and twelve input parameters were selected to develop the proposed ANN model. Optimal number of neurons and the percentage of data that should be used in the training and testing sets were determined for the ANN model. Three learning algorithms were considered for the neural network model and their performance were compared. The accuracy of optimized neural network was evaluated by experimental data. The results of the proposed ANN model were compared with the existing design codes and published studies. While the mean predicted-to-tested values of ANN model for lateral and flexural strengths were 1.019 and 1.038, respectively, the results of existing procedures had relatively inaccurate estimations ranged from 0.948 to 1.808.

P-1646. Assessing Antibiotic Adverse Drug Events from Walk-In Clinics at an Academic Healthcare System

Abstract Background Outpatient antibiotic adverse drug event (ADE) rates are largely unknown. This project aimed to evaluate methods for collecting data on antibiotic ADEs and report antibiotic ADE data from 8 walk-in clinics (WICs) at Vanderbilt University Medical Center (VUMC), which represents a high-volume antibiotic prescribing outpatient setting. Methods We attempted retrospective data collection of antibiotic ADE billing codes. After multiple trials, we recognized a need to understand how ADEs are documented (Tables 1 and 2) versus billed. To examine this, we conducted a retrospective chart review of patients prescribed antibiotics in selected WICs from 8/1/23 - 10/31/23 with WIC, primary care, or emergency department visits within 14 days after. We reviewed visit notes for antibiotic ADEs. If present, visit billing codes were collected and analyzed. To evaluate potential new antibiotic ADE data collection methods, especially ADEs reported via calls without a visit, we conducted prospective collection of antibiotic ADE calls or return visits within 14 days from patients prescribed antibiotics at selected WICs from 11/1/23-4/30/24 via voluntary REDCap survey reports by WIC providers. We used manual chart review to assess antibiotic appropriateness. Results From 8/1/23-10/31/23, we extracted 1059 unique visits with antibiotic prescriptions; 69 (6.5%) had antibiotic ADEs documented in subsequent visits. ADEs were secondary complaints in 30/69 (43.5%) visits. ADE billing codes were present in 17/69 (24.6%) visits (Figure 1). From 11/1/2023-4/30/2024, 16,274 WIC visits had antibiotic prescriptions. Providers logged 45 calls and 3 return visits (48/16274, 0.3%); 6/48 (12.5%) antibiotics were unnecessary. When antibiotics were indicated, choice, dose, or duration was inappropriate in 13/42 (31.0%).Table 2:ICD-10 codes classified by organ system Conclusion This project showed outpatient antibiotic ADEs occur routinely. Utilizing voluntary reporting led to increased capture of antibiotic ADE patient calls but was not comprehensive. Additionally, ADE billing codes are currently underutilized, resulting in inaccurate estimates of antibiotic ADEs. Future studies are needed to determine an optimal approach to identify and extract antibiotic ADE data from electronic medical records. Disclosures Milner Staub, MD, MPH, Gilead: Stocks/Bonds (Public Company)|Johnson & Johnson: Stocks/Bonds (Public Company)

Applying auxiliary supervised depth-assisted transformer and cross modal attention fusion in monocular 3D object detection.

Monocular 3D object detection is the most widely applied and challenging solution for autonomous driving, due to 2D images lacking 3D information. Existing methods are limited by inaccurate depth estimations by inequivalent supervised targets. The use of both depth and visual features also faces problems of heterogeneous fusion. In this article, we propose Depth Detection Transformer (Depth-DETR), applying auxiliary supervised depth-assisted transformer and cross modal attention fusion in monocular 3D object detection. Depth-DETR introduces two additional depth encoders besides the visual encoder. Two depth encoders are supervised by ground truth depth and bounding box respectively, working independently to complement each other's limitations and predicting more accurate target distances. Furthermore, Depth-DETR employs cross modal attention mechanisms to effectively fuse three different features. A parallel structure of two cross modal transformer is applied to fuse two depth features with visual features. Avoiding early fusion between two depth features enhances the final fused feature for better feature representations. Through multiple experimental validations, the Depth-DETR model has achieved highly competitive results in the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) dataset, with an AP score of 17.49, representing its outstanding performance in 3D object detection.

Enhanced Terrain-Referenced Navigation Through Adaptive Radar Altimeter Error Estimation with Monte Carlo Sampling

Abstract Terrain-referenced navigation serves as a reliable backup system for military aviation, but its performance is often compromised by inaccurate radar altimeter noise estimation. The wide beamwidth of radar altimeters causes complex error characteristics that vary with flight conditions and radar–terrain interactions. A mismatch between assumed and actual measurement error levels can lead to degraded navigation accuracy and potential filter divergence. This issue is exacerbated in mission scenarios such as low-altitude penetrations that require frequent rolling and pitching maneuvers. This study proposes a novel radar altimeter error modeling approach that uses Monte Carlo sampling to estimate measurement error levels on-the-fly. By incorporating a radar–terrain interaction measurement model, the method provides tighter error bounds to an extended Kalman filter, improving its response to state- and terrain-dependent error characteristics. Simulation across four distinct scenarios show that the proposed approach reduces the average divergence rate from 2.75 to 0.25%, and decreases horizontal position error by 32.3% in RMSE and 31.7% in CEP.

Co-Financing of Capital Investment Within the Frames of Federal Projects

The article deals with issues of inter-budget and inter-departmental coordination to finance project activities connected with capital investment. Special attention was paid to evaluation of factors influencing budget expenses of all levels and non-budget sources, observation of dates of reaching control points and results of project events in conducting capital investment. The goal of the present research is to identify problems in mechanism of capital investment financing within the frames of federal and regional projects at different levels of management and work out recommendations aimed at their overcoming. The authors revealed and systematized key problems in implementation of federal and regional projects in case of capital investment financing, analyzed their reasons and after-effects illustrated by different projects, including incompleteness of regulating enactments in the field of project financing of budget expenses; low quality of project-estimate documents; improper procedures of finding the supplier (contractor, executor) and careless observation of their obligations under the contract. Special attention was paid to issues of insufficient finance support of capital investment, inaccurate estimation of finance volumes and coordination of dates of reaching control points in project events. As a result the authors formulated and substantiated practical recommendations aimed at raising the role and wider authorities of municipal bodies as participants of project activities, upgrading procedures of project documents preparation, modernization and developing new information systems in the field of project activities of power bodies, elaborating typical provisions for cooperation of authorized bodies and other participants of purchase, managing risks of circumstances not under control of contract client and executors. These recommendations could provide more efficient planning of budget allocations in developing federal and regional projects, cutting terms of step implementation for building projects of capital construction, tracing the process of capital investment, minimizing risks of unpredictable costs of consolidated budgets of entities of the Russian Federation.

Remaining useful life prediction of lithium-ion batteries using a novel particle flow filter framework with grey model

Remaining useful life (RUL) prediction is a crucial aspect of the prognostics health management of lithium-ion batteries (LIBs). Owing to the influence of resampling technology, particle degradation is often observed in the particle filter-based RUL prediction of LIBs, resulting in a low prediction accuracy and large uncertainty. In this paper, a novel particle flow filter with the grey model method (GM-PFF) is proposed to forecast the RUL and state of health of batteries. First, the least squares method is employed to obtain the initial values for double exponential empirical model parameters. Subsequently, the grey model is used to predict the current cycle capacity of LIBs as an observation value for the particle flow filter, solving the inaccurate estimation problem of the state of particle flow filter observation values, and the particle flow filter method is employed to update model parameters. Finally, a test dataset is divided into early, middle, and late stages to predict the RUL of LIBs and obtain the probability distributions. On the CALCE and NASA PCoE LIB dataset, GM-PFF reduces RMSE by 1% compared to PFF, exhibiting a higher prediction accuracy and effectively addressing the particle degradation problem.

Artificial intelligence in four-dimensional imaging for motion management in radiation therapy

Four-dimensional imaging (4D-imaging) plays a critical role in achieving precise motion management in radiation therapy. However, challenges remain in 4D-imaging such as a long imaging time, suboptimal image quality, and inaccurate motion estimation. With the tremendous success of artificial intelligence (AI) in the image domain, particularly deep learning, there is great potential in overcoming these challenges and improving the accuracy and efficiency of 4D-imaging without the need for hardware modifications. In this review, we provide a comprehensive overview of how these AI-based methods could drive the evolution of 4D-imaging for motion management. We discuss the inherent issues associated with multiple 4D modalities and explore the current research progress of AI in 4D-imaging. Furthermore, we delve into the unresolved challenges and limitations in 4D-imaging and provide insights into the future direction of this field.

Robust Kalman Filter with Recursive Measurement Noise Covariance Estimation Against Measurement Faults

A new innovation-based recursive measurement noise covariance estimation method is proposed. The presented algorithm is used for Kalman filter tuning, as a result, the robust Kalman filter (RKF) against measurement malfunctions is derived. The proposed innovation-based RKF with recursive estimation of measurement noise covariance is applied for the model of Unmanned Aerial Vehicle (UAV) dynamics. Algorithms are examined for two types of measurement fault scenarios; constant bias at measurements (additive sensor faults) and measurement noise increments (multiplicative sensor faults). The simulation results show that the proposed RKF can accurately estimate UAV dynamics in real time in the presence of various types of sensor faults. Estimation accuracies of the proposed RKF and conventional KF are investigated and compared. In all investigated sensor fault sceneries, the Root Mean Square (RMS) errors of the proposed RKF estimates are lower. The conventional KF gives inaccurate estimation results in the presence of sensor faults.

Flexible extended warranty considering products’ heterogeneous reliability

ABSTRACT Customers often have the option to extend the warranty of a new durable product by paying extra for an extended warranty (EW). A critical issue in EW design is EW cost estimation, which highly depends on the field reliability of the product. An inaccurate warranty cost estimation may develop inappropriate risk management leading to high or low financial capacity with warranty claims and inefficiency on customer service. In practice, the field reliability of sold items differs due to variations in manufacturing process, usage, and operating conditions. However, this heterogeneity is not considered in traditional uniform warranty strategies. To recognize the reliability of the item, we apply Bayesian inference to the lifetime model with a given prior based on the observed failure data. Moreover, we propose a flexible EW strategy allowing delayed EW purchases within the base warranty period by considering the EW cost estimation based on the field reliability. Specifically, we consider different preventive maintenance (PM) strategies based on the updated reliability model and analyze the EW costs for different scenarios. Numerical experiments and actual examples of diesel engines are presented to validate the proposed method.

Adjustment of Truncation Effect in First Birth Interval using Current Status Data Technique

Introduction: Estimating the First Birth Interval (FBI) from cross-sectional data often presents challenges related to truncation effects. These challenges stem from the data’s inability to capture the enough exposure for an event, resulting in potential biases and inaccuracies in FBI estimates. Recognizing and addressing truncation effects is essential for obtaining more precise and meaningful fertility parameter estimates in a cross-sectional survey. This study seeks to mitigate truncation effects in the estimation of the FBI by utilizing the Current Status Data technique. This approach focuses on women with specific marital durations, providing a means to counteract the bias caused by truncation and thereby yielding more accurate and reliable FBI estimates. Methods: Data from the National Family Health Survey (NFHS-IV) are employed for this study. The Current Status Data Technique is applied to the dataset, considering exclusively those women with marital durations less than 120 months. This methodology enables the adjustment of truncation effects and facilitates a more precise estimation of the FBI. Statistical analysis is conducted to determine the FBI distribution and ascertain the necessary sample size. Results: The estimated First Birth Interval (FBI) without accounting for truncation is 27.85 months, while the estimate considering truncation is 31.70 months. When applying the Current Status Data technique, the estimated FBI is 30.70 months. To obtain reliable estimates of the FBI using Current Status techniques, a minimum sample size of over 5,000 observations is necessary. Conclusion: The truncation effect in FBI is addressed, and some non-parametric adjustments are used for estimating the duration of FBI. The Current Status Data technique emerges as a valuable tool for mitigating these effects and enhancing the precision of FBI estimates. This research contributes to an improved understanding of fertility dynamics and provides valuable insights for future studies on the First Birth Interva

A Score Function to Prioritize Editing in Household Survey Data: A Machine Learning Approach

Errors in household finance survey data collection can lead to inaccuracies in population estimates. Manual case-by-case revision has traditionally been used to identify and edit potential errors and omissions in the data, such as omitted or misreported assets, income, and debts. Selective editing strategies aim at reducing the editing burden by prioritizing cases through a scoring function. However, the application of traditional selective editing strategies to household finance survey data is challenging due to their underlying assumptions. Using data from the Spanish Survey of Household Finances, we develop a machine learning approach to classify data during the editing phase into cases affected by severe errors and omissions. We compare the performance of several supervised classification algorithms and find that a Gradient Boosting Trees classifier outperforms the competitors. We then use the resulting score to prioritize cases and consider data editing efforts into the choice of an optimal classification threshold.

Degradation-aware dynamic kernel for blind super-resolution

ABSTRACT Image super-resolution (SR) methods based on convolutional neural networks have become mainstream. However, unknown degradation in test images significantly affects SR performance. Existing blind SR methods struggle with inaccurate kernel estimation in explicit modeling and over-reliance on implicit degradation information. To address these issues, we propose the Degradation-Aware Dynamic Kernel with Dual-path Attention for Super-Resolution (DADKSR) network. DADKSR employs degradation representation learning to enable adaptive kernel estimation and uses a dual-path attention module to enhance spatial and channel features. An enhanced detail feature fusion (EDFF) module further enriches texture details, while the cycle correction learning strategy utilizes the SR results with the real degradation kernel in the low-resolution (LR) image to further correct the degradation kernel. Experimental results demonstrate that DADKSR improves kernel estimation accuracy, achieving an average PSNR improvement of 0.76 dB on public datasets with anisotropic Gaussian kernels, producing clearer structures and edges.

International Expert-Based Consensus Definition, Classification Criteria, and Minimum Data Elements for Osteoradionecrosis of the Jaw: An Inter-Disciplinary Modified Delphi Study.

Osteoradionecrosis of the jaw (ORNJ) is a severe iatrogenic disease characterized by bone death after radiation therapy (RT) to the head and neck. With over 9 published definitions and at least 16 classification systems, the true incidence and severity of ORNJ are obscured by lack of a standard for disease definition and severity assessment, leading to inaccurate estimation of incidence, reporting ambiguity, and likely under-diagnosis worldwide. This study aimed to achieve consensus on an explicit definition and phenotype of ORNJ and related precursor states through data standardization to facilitate effective diagnosis, monitoring, and multidisciplinary management of ORNJ. The ORAL Consortium comprised 69 international experts, including representatives from medical, surgical, radiation oncology, and oral/dental disciplines. Using a web-based modified Delphi technique, panelists classified descriptive cases using existing classification systems, reviewed systems for feature extraction and specification, and iteratively classified cases based on clinical/imaging feature combinations. The Consortium ORNJ definition was developed in alignment with SNOMED-CT terminology and recent ISOO-MASCC-ASCO guideline recommendations. Case review using existing ORNJ classification systems showed high rates of inability to classify (up to 76%). Ten consensus statements and nine minimum data elements (MDEs) were outlined for prospective collection and classification of precursor/ORNJ disease states. This study provides an international, consensus-based definition and MDE foundation for standardized ORNJ reporting in cancer survivors treated with RT. Head and neck surgeons, radiation, surgical, medical oncologists, and dental specialists should adopt MDEs to enable scalable health information exchange and analytics. Work is underway to develop both a human- and machine-readable knowledge representation for ORNJ (i.e., ontology) and multidisciplinary resources for dissemination to improve ORNJ reporting in academic and community practice settings.

A GNN-Based Placement Optimization Guidance Framework by Physical and Timing Prediction

Placement is crucial in physical design flow with significant impact on later routability and ultimate manufacturability in terms of performance, power, and area (PPA), which may deviate from finding the optimal solution and/or lead to unnecessary iterations suffering from interleaved optimization steps and inaccurate PPA estimation. To solve this issue, we propose a physical- and timing-related placement optimization guidance framework which provides candidate gate sizing and buffer insertion solutions as well as a path group for potential violated paths based on graph neural networks (GNNs) to improve placement quality significantly and efficiently. Experimental results on the OpenCores benchmarks with 22 nm technology demonstrate that the proposed placement optimization guidance framework achieves up to 35.66% and 43.51% worst negative slack (WNS) and total negative slack (TNS) improvement and 52.17% reduction in the number of violating paths (NVP), which is beneficial to later routing stages with 2.33% wirelength decrease.

Risk Stratification for Cardiovascular Disease: A Comparative Analysis of Cluster Analysis and Traditional Prediction Models.

Primary prevention of cardiovascular disease (CVD) relies on effective risk stratification to guide interventions. Current models, primarily developed using regression analysis, can lead to inaccurate estimates when applied to external populations. This study evaluates the utility of cluster analysis as an alternative method for developing CVD risk stratification models, comparing its performance with established CVD risk prediction models. Using data from 3,416 individuals (mean age of 66 years and no prior CVD) followed for an average of 5.2 years for incidence of CVD, we developed a risk stratification model using cluster analysis based on established CVD risk factors. We compared our model to the Systematic Coronary Risk Evaluation (SCORE2), the Pooled Cohort Equations (PCE) and the Predicting Risk of Cardiovascular Disease Events (PREVENT) models. We used Poisson and Cox regression to compare CVD risk between risk categories in each model. Predictive accuracy of the models was evaluated using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and C-statistic. During the study, 161 CVD events were detected. The high-risk cluster had a sensitivity of 59.0%, a PPV of 7.5% a specificity of 64.2% and NPV of 96.9% to predict CVD. Compared to the high-risk groups of the SCORE2, PCE and PREVENT, the high-risk cluster had a high sensitivity and NPV, but a low specificity and PPV. No statistically significant differences were found in C-statistic between models. Cluster analysis performed comparably to existing models and identified a larger high-risk group that included more individuals who developed CVD, though with more false positives. Further studies in larger, diverse cohorts are needed to validate the clinical utility of cluster analysis in CVD risk stratification.

NHTSA crash investigation sampling system (CISS) is unreliable, inaccurate and does not estimate serious injury in motor vehicle crashes.

Errors and inaccuracies are documented in the 2017-2022 CISS (Crash Investigation Sampling System) field data managed by NHTSA on serious injury in motor vehicle crashes in the US. All CISS cases of serious, non-fatal injury to 3-5 and 6-12 year old (yo) children were downloaded from the NHTSA website. Each case was summarized for the: 1) case identification by year, PSU and number, 2) case weighting factor and domain, 3) vehicle make, model and model year and 4) age, seating position and maximum injury severity (MAIS) to the child. Crashes with serious injury are in domains 02, 05 and 08 depending on the age of the vehicle. There were several cases with unusually high weighting factors (casewgt) in comparison to others in the groups. The sources of the high case weights were identified. There were 8 cases with MAIS 3-6 injury to 3-5 yo children. The weighted sum was 3,538.8 estimated national incidence for 6 years. Case #5 had a weight of 2,223.3, which was 63% of the sum. There were 8 cases with MAIS 4-6 injury to 6-12 yo children. The weighted sum was 5,817.9. Case #9 had a weight of 5,199.6, which was 89% of the sum. There were 18 cases with MAIS 3 injury to 6-12 yo children. Case #22 had a weight of 1,264.5 and case #25 937.8, representing 52% of the sum. The high case weights were traced to errors in the domain due to incorrect injury severity and vehicle age. NHTSA has been aware of the domain errors and unusually high case weights since 2012. They have not corrected the errors in high case weights. They have ignored the effects of errors in the CISS and NASS-CDS data on serious injury. The data errors lead to unrealistic and inaccurate estimates of serious injury. There are other problems with CISS and NASS-CDS. There are pervasive errors in the CISS crash selection procedure, domains and case weights that make estimates of serious injury unreliable. There is no quality control to correct the data errors. CISS cases are incomplete and limited in comparison to the earlier NASS-CDS.CISS sampling procedure is incorrect by relying on vehicle age for case selection. NASS-CDS sampling relied on occupant injury severity. CISS and NASS-CDS incorrectly handle multi-vehicle crashes. A complete review and revision of CISS and NASS-CDS is needed to correct errors so the databases are compatible from the 1980s to today. NHTSA needs to disclose information about the revised NASS-CDS and CISS files. The online NASS-CDS and CISS files are incorrect and inadequate to estimate serious injury in motor vehicle crashes.CISS sampling frequencies rely on 2011 NASS-CDS, which is not relevant to serious injury crashes today. The 2011 NASS-CDS did not investigate older vehicles. CISS under-samples serious injury in older vehicles where race, gender and SES are important factors. CISS and NASS-CDS are not representative samples of serious injury in US crashes.

Composite Learning Based Adaptive Control of Linear 2×2 Hyperbolic PDE Systems.

This article considers the adaptive stability control of a class of 2×2 linear hyperbolic PDE systems. The PDE model is subject to constant but in-domain and boundary unknown parameters. A novel adaptive controller is developed by leveraging the swapping design technique and composite parameter learning law. With swapping design, several linear and static combinations, including carefully designed filters, unknown parameters, and error terms, are constructed to express the system states. From the static combinations, a composite learning based forgetting-factor least squares law is introduced to guarantee exponential parameter convergence without the persistent excitation (PE). Although inaccurate parameter estimation in the adaptive backstepping control results in asymptotic stability of the system, accurate parameter estimation ensures the exponential convergence of closed-loop system and concomitantly improves the transient performance. Finally, a comparative numerical simulation is performed to validate the effectiveness and advantage of the developed adaptive control scheme.

Summer CH4 ebullition strongly determines year-round greenhouse gas emissions from agricultural ditches despite frequent dredging.

Recent studies indicate that greenhouse gas (GHG) emissions from agricultural drainage ditches can be significant on a per-unit area basis, but spatiotemporal investigations are still limited. Additionally, the impact of dredging - a common management in such environments - on ditch GHG emissions is largely unknown. This study presents year-round GHG emissions from nine ditches on a dairy farm in the center of the Netherlands, where each year, approximately half of the ditches are dredged in alternating cycles. We measured monthly diffusive fluxes of carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O), along with ebullitive CH4 emissions, supplemented by diel flux measurements (i.e., 24-h measurements) conducted in summer and winter. Our findings indicate that while diffusive GHG fluxes exhibited low spatiotemporal variation, ebullitive CH4 emissions were significantly higher during warmer periods and marginally elevated at ditch intersections. CH4 ebullition was the dominant pathway of ditch GHG emissions, accounting for 58% of the total annual emissions, followed by CO2 (39%), and N2O (3%). Approximately 80% of the total CH4 emissions occurred through ebullition during spring and summer. The average CH4 emission factor estimated for our ditches (574kgha-1 year-1) is ∼40% higher than the Tier-1 value suggested by the IPCC for ditches on mineral soils (416kgha-1 year-1). Based on two 24-h measurement campaigns, we concluded that neglecting nighttime diffusive CO2 and CH4 emissions may lead to inaccurate estimates of annual ditch GHG emissions, with ∼12% underestimation in this study. Although dredging led to subtle increases in water-to-atmosphere GHG emissions immediately after the activity, it reduced overall annual GHG emissions by ∼35%. This study highlights the importance of CH4 ebullition and of capturing diel cycles of diffusive emissions to accurately assess ditch GHG emissions and underscores the importance of considering seasonal variations and dredging practices when budgeting ditch GHG emissions.

PENGARUH MODEL PEMBELAJARAN DAN KEGIATAN PRAKTIKUM TERHADAP PEMAHAMAN KONSEP PESERTA

This study aims to determine the influence of learning models and practicum activities on students' conceptual understanding, as well as to examine the interaction between learning models and practicum activities on students' conceptual understanding. The research was conducted using a quantitative descriptive method with a 3x2 factorial experimental design, and the data were analyzed using two-way ANOVA to identify differences in variance between two treatment groups. Based on the analysis results, it can be concluded that there is an effect of learning models on students' conceptual understanding. However, there is no effect of conducting practicum activities either at the beginning or at the end of the learning process on students' conceptual understanding. Additionally, there is no interaction between learning models and practicum activities, as practicum activities are considered part of the learning model. The discrepancy between the results and theoretical expectations is influenced by several factors, including students' initial understanding and skills in conducting practicum activities. Furthermore, inaccurate time estimation in executing each step of the learning syntax also contributed to these findings.