Importance Weights Research Articles

Prognostic prediction of cardiovascular adverse events in patients after percutaneous coronary intervention using machine learning

Abstract Introduction While prevention of major cardiovascular adverse events (MACE) is important in the management of patients after percutaneous coronary intervention (PCI), risk stratification of MACE remains a significant challenge with increasingly diverse and complex patient backgrounds. In response to that, this retrospective observational study aimed to explore whether machine learning (ML) methods can predict future MACE after PCI using a variety of features extracted from electronic medical records (EMRs). Methods In 3,016 patients with coronary artery disease who performed PCI, 3 machine learning (ML) algorithms (gentle boost, random forest, and neural network) and statistical model were retrospectively applied to predict the occurrence of MACE within two years after PCI (564 patients, 18.7%). MACE was defined as composite endpoint of cardiac death, myocardial infarction (MI), any revascularization, and congestive heart failure (CHF). Basic patient information (e.g., age, sex, coronary risk factors) and laboratory test results were used for the analysis. Feature importance was calculated to identify key factors for MACE. Hold-out method with an 8:2 split between training and test was used to calculate sensitivity, specificity, positive predictive value, negative predictive value, and F-value for each model. Results As shown in Figure 1, ML models showed excellent diagnostic accuracies in the prediction of MACE after PCI, demonstrating higher F-values compared with statistical model (logistic regression analysis). Similar results were seen in the subgroup analyses for estimating cardiac death, MI, revascularization, or CHF. Although the weighting of feature importance for MACE recurrence varied among ML models, infarct size (peak CK levels), increased inflammatory status, and traditional coronary risk factors were identified as important features for predicting MACE within two years after PCI (Figure 2). Conclusions It is suggested that ML using EMRs is promising for inferring future MACE within two years after PCI. Further studies are warranted to address whether ML-based risk stratification could lead to improved clinical decision-making after PCI.

A simulation study towards sparse-view spectral X-ray CT segmentation using spectrum-aware deep neural networks

ABSTRACT Spectral X-ray computed tomography (SXCT) is able to provide richer spectral information than conventional X-ray CT, thereby improving the capability of discriminating materials in segmentation tasks. However, segmentation algorithms for conventional CT images often lose their performance in material segmentation with SXCT images due to the spectrum-dependent attenuation properties of X-rays. Recently, deep neural networks (DNNs) have demonstrated their effectiveness in segmentation tasks. However, processing SXCT with DNNs requires significant memory resources and may take significant computation time. To address these issues, this paper introduces a simulation study towards realising deep-learning-based SXCT image segmentation. The proposed method employs a convolutional neural network (CNN) that selectively utilises the significant spectrum channels for accurate material segmentation. The proposed method builds upon a prior study on adaptive CT reconstruction, but to further enhance the performance, we introduce a self-attention-based spectrum-selection module. During training, the selection module assigns importance weights to spectrum channels based on the attention scores. Once the network has been trained, the proposed method can obtain satisfactory segmentation results from projection images at one or a combination of several spectrum channels recommended by the selection module. The experimental results demonstrate that the proposed method achieves a significant reduction in computation time while maintaining the segmentation quality.

MEMPERLUAS KEMAMPUAN ALAT PENCABUT SINGKONG: PENGEMBANGAN LEBIH LANJUT UNTUK KEBUTUHAN PETANI

This research builds upon the author's earlier work on the design of a cassava pulling tool. The aim is to develop a non-mechanical or minimally-powered tool that addresses farmer needs. The tool design incorporates Quality Function Deployment (QFD) to ensure user satisfaction. QFD is a method that integrates customer desires into technical requirements during the design phase, prior to production. Data analysis revealed key user desires: lightweight materials, affordability, the ability to pull multiple cassava stems simultaneously, durability, safety, and additional functionalities. House of Quality (HOQ) analysis identified the most crucial attribute (weight of importance: 30%) as the ability to pull multiple stems at once. This finding served as the primary basis for prototype development and technical characteristic definition. Subsequent prototype testing successfully uprooted cassava plants, demonstrating the tool's capability of pulling multiple stems in a single operation.

Comprehensive chromatographic profiling and structural analysis of key anticoagulant components in enoxaparin

Heparin is the most widely used anticoagulant in clinical practice, with enoxaparin being one of the most important low molecular weight heparins (LMWHs). In this study, an antithrombin III (ATIII) affinity column was used. Enoxaparin and its oligosaccharides of varying sizes, prepared using preparative size exclusion chromatography (SEC), were fractionated through the ATIII affinity column. The different affinity fractions from each oligosaccharide size were profiled using strong anion exchange (SAX) chromatography. Each peak was automatically transferred to an SEC column for desalting prior to mass spectrometry (MS) analysis, which enabled structural identification using a multiple heart-cut (MHC) 2D LC-MS system (SAX-SEC-MS). The high-affinity fraction from enoxaparin was further analyzed using the MHC 2D LC system (SEC-SAX). SAX profiles of the high-affinity oligosaccharides, prepared by both size and affinity fractionation, were consistent with those obtained by direct SEC-SAX analysis. The possible sequences of several high-affinity hexasaccharides and the domain compositions of high-affinity octa- and decasaccharides in enoxaparin were further elucidated by disaccharide analysis after manual collection of the oligosaccharides. This work advances the understanding of enoxaparin's structural features and offers a potential approach to improve the quality of enoxaparin, as well as to identify key structural motifs in heparin/LMWHs that contribute to protein binding.

Integrated and sustainable performance evaluation of urban rail transit systems using fuzzy sustainability index

The complexity of prominent dimensions in urban rail transportation sector arises from the involvement of several stakeholders and the various interacting variables for evaluating transportation performance. This study uses a fuzzy logic-based method to present an index-based framework for sustainable performance assessment. The proposed framework presents 4 main indicators which include environmental, economic, financial and social sustainability, 14 sub-indicators, 4 main indicators for transportation system effectiveness with 13 sub-indicators and 1 main indicator for multi modal integration with 6 sub-indicators. The performance ratings and importance weights of the transport sustainability attributes are evaluated using a fuzzy logic approach. A numerical illustration is provided to demonstrate the applicability of the approach. The results show that system capacity (passenger capacity of a route), comfort and customer service (passenger environment and ride quality, maintenance cycle) and multimodal integration (park and ride facilities) were found to be weaker attributes. The proposed methodology facilitates the systematic understanding of the prominent attributes for achieving integrated and sustainable transport performance by using the ‘Fuzzy Transport Sustainability Index’.

A lightweight defect detection algorithm for escalator steps

In this paper, we propose an efficient target detection algorithm, ASF-Sim-YOLO, to address issues encountered in escalator step defect detection, such as an excessive number of parameters in the detection network model, poor adaptability, and difficulties in real-time processing of video streams. Firstly, to address the characteristics of escalator step defects, we designed the ASF-Sim-P2 structure to improve the detection accuracy of small targets, such as step defects. Additionally, we incorporated the SimAM (Similarity-based Attention Mechanism) by combining SimAM with SPPF (Spatial Pyramid Pooling-Fast) to enhance the model’s ability to capture key information by assigning importance weights to each pixel. Furthermore, to address the challenge posed by the small size of step defects, we replaced the traditional CIoU (Complete-Intersection-over-Union) loss function with NWD (Normalized Wasserstein Distance), which alleviated the problem of defect missing. Finally, to meet the deployment requirements of mobile devices, we performed channel pruning on the model. The experimental results showed that the improved ASF-Sim-YOLO model achieved an average accuracy (mAP50) of 96.8% on the test data set, which was a 22.1% improvement in accuracy compared to the baseline model. Meanwhile, the computational complexity (in GFLOPS) of the model was reduced to a quarter of that of the baseline model, while the frame rate (FPS) was improved to 575.1. Compared with YOLOv3-tiny, YOLOv5s, YOLOv8s, Faster-RCNN, TOOD, RTMDET and other deep learning-based target recognition algorithms, ASF-Sim-YOLO has better detection accuracy and real-time processing capability. These results demonstrate that ASF-Sim-YOLO effectively balances lightweight design and performance improvement, making it highly suitable for real-time detection of step defects, which can meet the demands of escalator inspection operations.

Smart retrofitting for office buildings: Comparison of decision-making criteria between developing and developed regions

Given the constantly evolving nature of modern infrastructure, retrofitting existing buildings with state-of-the-art smart technologies presents a pivotal opportunity, enabling promising advancements in sustainability, efficiency and adaptability. Focusing on Hong Kong and Sri Lanka, which are developed and developing regions respectively, this study investigates the decision-making criteria for selecting smart systems as retrofits for office buildings. Using the analytic network process (ANP), 24 expert interviews were conducted in Hong Kong and Sri Lanka to determine the importance weights of the decision-making criteria. The findings show varied importance weights for each criterion between the two regions, with Hong Kong emphasizing upfront financial considerations, specifically 'investment cost' while Sri Lanka emphasizing long-term operational sustainability, specifically 'recurring operation, maintenance and repair cost.' Despite these differences, there are similarities in the importance of core criteria such as 'reliability of retrofit option' and 'ability to increase returns.' The lack of focus on user-centric issues in Hong Kong and implementation of retrofit projects in Sri Lanka underscore the need for decision-makers to prioritize these aspects in future projects. The principal contribution of this study is in offering insights from a thorough analysis of the importance weights of the essential criteria in making decisions regarding smart retrofitting (SR). It is also beneficial to various stakeholders, including building owners and facility managers. The cross-region study approach, enabling the investigation on both developed and developing regions, further serves as a foundation for future research aiming to explore and analyze the influence of regional difference in decision-making for SR.

How Much Better is Faster? Empirical Tests of QALY Assumptions in Health-Outcome Sequences.

This study was designed to test hypotheses regarding the path dependence of health-outcome values in the form of linear additivity of health-state utilities and diminishing marginal utility of health outcomes. We employed a discrete-choice experiment to quantify patient treatment preferences for major depressive disorder. In a series of choice questions, participants evaluated seven symptom-improvement sequences and out-of-pocket costs over 6-week durations. Money-equivalent values were derived from a deductive latent-class mixed-logit analysis. The discrete-choice experiment was completed by 751 respondents with self-reported major depressive disorder recruited from an online commercial panel. The class-membership probability was 0.83 for latent-class preferences consistent with supporting relative importance weights for all symptom-improvement sequences in the study design. First, we found strong support for diminishing marginal utility in symptom-improvement sequences. The money-equivalent value of an initial week of normal mood was $147 (95% confidence interval: $128, $166) and a second week of normal mood was $70 ($49, $91). Furthermore, for short treatment durations where conventional discounting was not a factor, equivalent changes in health status were valued more highly for an earlier onset of effect: holding subsequent symptom patterns constant, $338 (211, 454) versus $70 (49, 91) for improvements starting in week 2 versus week 3 and $147 ($128, $166) versus $29 (-$4, $64) for improvements starting in week 3 versus week 4. Our findings imply that conventional quality-adjusted life-year calculations in which health values are assumed to be path independent can understate the value of health improvements that appear earlier in a sequence.

What do you want to eat? Cuisine and nutrition intervention preferences among people using a large food pantry in Texas

ObjectiveNutrition interventions delivered through food pantries could reduce health disparities for people experiencing food insecurity. We identified clients’ preferences for cuisines, nutrition interventions, and outcomes and whether preferences differ for subpopulations. MethodsCross-sectional study at a large pantry in Dallas, Texas (N = 200). Survey collected from February-May 2023 on demographics, cuisine preferences, nutrition intervention preferences, and outcomes clients hope to achieve when changing lifestyle (weight loss, feeling comfortable in clothes, feeling good about diet, wellbeing). A subsample (N = 130) had height and weight measured. We tested whether food security and BMI (categorical) were associated with intervention or outcome preferences using IBM SPSS Statistics (Version 29) to conduct analysis of variance. ResultsTop-rated cuisines were Mexican, Chinese, Italian. Participants reported a desire for interventions implemented through the pantry reflected by high Nutrition Intervention Index scores. The highest rated intervention was bringing more healthy food into the pantry and lowest rated was restricting unhealthy donations.Overall wellbeing was the most important outcome and weight loss the least important.Neither food security nor BMI were associated with desire for interventions. All outcomes were rated in a similar pattern, though people with obesity and overweight rated weight loss as more important than people with normal weight. ConclusionsMost participants demonstrated a strong desire for healthier, ethnically diverse options, and nutrition interventions delivered through the pantry. Our findings explore cuisines and outcomes preferred by people that use food pantries which can guide researchers, clinicians, and non-profit organizations in planning and promotion of nutrition programs for pantry clients.

An Iterative Method for Unsupervised Robust Anomaly Detection Under Data Contamination.

Most deep anomaly detection models are based on learning normality from datasets due to the difficulty of defining abnormality by its diverse and inconsistent nature. Therefore, it has been a common practice to learn normality under the assumption that anomalous data are absent in a training dataset, which we call normality assumption. However, in practice, the normality assumption is often violated due to the nature of real data distributions that includes anomalous tails, i.e., a contaminated dataset. Thereby, the gap between the assumption and actual training data affects detrimentally in learning of an anomaly detection model. In this work, we propose a learning framework to reduce this gap and achieve better normality representation. Our key idea is to identify sample-wise normality and utilize it as an importance weight, which is updated iteratively during the training. Our framework is designed to be model-agnostic and hyperparameter insensitive so that it applies to a wide range of existing methods without careful parameter tuning. We apply our framework to three different representative approaches of deep anomaly detection that are classified into one-class classification-, probabilistic model-, and reconstruction-based approaches. In addition, we address the importance of a termination condition for iterative methods and propose a termination criterion inspired by the anomaly detection objective. We validate that our framework improves the robustness of the anomaly detection models under different levels of contamination ratios on five anomaly detection benchmark datasets and two image datasets. On various contaminated datasets, our framework improves the performance of three representative anomaly detection methods, measured by area under the ROC curve.

Remaining useful life prediction method for cross- condition tools based on parallel fusion

Abstract The current tool remaining useful life (RUL) prediction models do not fully consider the importance of signals from different sensors and the differences in working conditions, which leads to a decline in the model’s robust performance under new conditions. This paper introduces a novel cross-condition RUL prediction approach that integrates parallel fusion, transfer learning, and deep learning. Considering the characteristics of different data sources, correlated signals are initially selected using correlation analysis and dynamic time warping (DTW). Subsequently, multi-scale Depthwise Separable Convolution and multi-attention mechanisms are combined to capture the importance of features at various scales. Finally, a dynamic domain adaptation technique is employed to adjust the importance weights of conditional distribution discrepancy and marginal distribution discrepancy, enabling the model to adapt to cross-condition scenarios.

Enhancing readiness degree for Industrial Internet of Things adoption in manufacturing enterprises: An integrated Pythagorean fuzzy approach

This study proposes a hybrid multi-criteria decision-making (MCDM) methodology designed to enhance the readiness for adopting the Industrial Internet of Things (IIoT) in manufacturing enterprises. The Pythagorean fuzzy approach is employed to address uncertainty and imprecision throughout decision-making processes. The development framework in this study incorporates TOE (Technology-Organization-Environment) and HOT fit (Human-Organization-Technology) to pinpoint barriers to IIoT adoption. Additionally, a triple helix model (THM) emphasizing on the synergy among university-industry-government is utilized to formulate pragmatic strategies. The agro-food processing industry in Thailand is used as a case study. In this study, even barriers are identified and validated through the Delphi method. The SWARA (Step-wise Weight Assessment Ratio Analysis) method determines the importance weights of these barriers, revealing “lack of digital culture”, “lack of knowledge and expertise,” and “job displacement concerns” as the three most critical barriers. The COBRA (COmprehensive Distance Based Ranking) method is employed to prioritize pragmatic strategies under THM, indicating that the role of the university in enhancing human capital capabilities is the most important, followed by the government's roles in enabling national ICT infrastructures and offering investment incentives as the second and third pragmatic strategies, respectively. A sensitivity analysis validates the proposed framework's reliability and robustness. The study's findings emphasize the potential of this integrated framework to guide future research endeavors among scholars and academicians across diverse industries beyond agri-food processing.

Effects of Letrozole And Metformin Versus Metformin alone on incretin in Iraqi patients with PCOS

I Background: The emergence of numerous novel therapeutic agents for managing T2DM has expanded the assortment of tailored therapies available for PCOS sick. One illustration of these is the incretin-based therapeutic agents. Patients and methods:The extanttrainingcomprised 75 sickthrough an age extent of 18 to less than 40 years. Patients were categorized into three groups: metformin group (positive control group) and they received metformin as 500 mg per-oral (bid); letrozole group who were treated using 2.5 mg per-oral (bid), and combination groups who received both agents with similar doses as above. Each group included 25 females. Informationalmost age and body mass index (BMI) stayed included in the study. Serum measurement of incretin (GLP-1) was done before treatment and 90 days after treatment using Enzyme-Linked Immunosorbent Assay. Results: There stayed no importantvariance in mean age (p = 0.981) betweentraining groups and means of age were 29.24 ± 6.10 years, 29.56 ± 5.50 years and 29.44± 5.83 years, respectively and the range of age was between 18 and 40 years. In addition, in this training, there stayed no importantalteration in mean BMI (p = 0.534) among study groups and means of BMI were 28.80 ±2.72 kg/m2, 27.53 ±1.95 kg/m2 and 27.07 ±1.99 kg/m2, respectively and the range of BMI was between 21.07 and 31.18 kg/m2. Serum incretin (GLP-1) level showed no significant variation among study groups, the mean levels were in the range of 10.87 -13.79 ng/ml and the range of values was from 9.56 -14.49 ng/ml. Changes in BMI are shown in. Metformin alone caused in more important weight decrease; nevertheless, adding of letrozole caused significant reduction in body mass index, letrozole alone also affect body weight significantly. It was observed that combined use of either drug resulted in more significant change of GLP-1 level at (p ≤ 0.001). Conclusion: Combined treatment with letrozole and metformin is safe and efficient in femalesby PCOS resultant in weight decrease, improved insulin

Identifying the most critical side effects of antidepressant drugs: a new model proposal with quantum spherical fuzzy M-SWARA and DEMATEL techniques

Identifying and managing the most critical side effects encourages patients to take medications regularly and adhere to the course of treatment. Therefore, priority should be given to the more important ones, among these side effects. However, the number of studies that make a priority examination is limited. There is a need for a new study that determines which of these effects are more priority to increase the quality of the treatment. Accordingly, this study aims to define the most important side effects of antidepressant drugs with a novel model. Quantum Spherical fuzzy M-SWARA technique is considered to compute the importance weights of the items. The main contribution of this study is that the most critical side effects can be understood for antidepressant drugs by establishing a novel decision-making model. The findings demonstrate that psychological side effects are defined as the most critical side effects of antidepressant drugs. Furthermore, physical side effects also play a key role in this condition. Side effects in antidepressant treatment have a great impact on the effectiveness of treatment and patient compliance.

An inter-hospital performance assessment model for evaluating hospitals performing hip arthroplasty

The value of hospital care to patients is expressed as a combination of reduced healthcare costs, fewer medical complications, and improved patient satisfaction. Few studies highlight the value hospitals provide to their patients through hip replacement surgery.This study aims to define a methodology for inter-hospital comparison purposes that can assess the value of hip replacement management to patients by using indicators of costs, medical complications, and patient outcomes.We identified medical complications and costs from medico-administrative data collected by three hospitals. We associated a Disability Adjusted Life Years (DALYs) impact with medical complications, readmissions (within 30 days), and hospital mortality. Costs were analysed from a social security perspective. Patient outcomes were collected through a questionnaire-based survey after hip surgery. To compare the three hospitals, we created a composite indicator by standardizing each dependent variable and combining a weighting of importance provided by patients.This study analysed 342 hospital stays. The mean (standard deviation) number of DALYs per stay was estimated to be more than 0.0028 (0.016) for a mean (standard deviation) cost of €4,834 (€3,665). The composite indicator allowed hospitals to be ranked and areas for improvement to be identified. In our case mix, Hospital 3 is the lowest-ranked hospital, with excessively high costs and a relatively low level of satisfaction compared to the others.The simultaneous evaluation of medical complications, patient outcomes, and costs is a prerequisite for quality improvement efforts by managers and practitioners. In our opinion, this experiment, which sought to estimate the value hospitals bring to patients, may be viewed as the first step towards value-based purchasing in Belgium.

Reliability-based ordinal consensus adjustment model for large scale group decision making

In large scale group decision-making (LSGDM), there are the substantial number of decision makers (DMs) with diverse knowledge, backgrounds, and interests related to the decision-making problem, and it is not possible to assure that all DMs are completely reliable. Thus, in order to enhance the quality of decision-making, it is necessary to analyze the reliabilities of DMs in LSGDM. This paper proposes the method to evaluate the reliabilities of DMs, sorts these DMs according to their degree of reliability, and investigates the consensus reaching process based on categories and an ordinal consensus measure. Considering the DMs' trust network, the uncertainty of a DM's evaluation information represented by a fuzzy preference relation (FPR), the deviation between a DM's FPR and those of the other DMs, and additive consistency of FPRs, the reliability of a DM is assessed using four criteria: PageRank centrality, professional competence, collaborative competence, and additive consistency. Following these reliability assessment criteria, ELECTRE-TRI is employed to sort DMs into three ordered categories according to DMs' different levels of reliability under the four assessment criteria. Furthermore, an improved ordinal consensus measure is designed to consider both the importance weights of positions and the deviation of Borda counts of the same alternative in two rankings. As for the consensus reaching process, due to the varied reliabilities of DMs in different categories, we propose a multiple strategies feedback mechanism for DMs in different categories. Finally, a numerical example is provided to illustrate the rationality and validity of the proposed model.

An Iteratively Reweighted Importance Kernel Bayesian Filtering Approach for High-Dimensional Data Processing

This paper proposes an iteratively re-weighted importance kernel Bayes filter (IRe-KBF) method for handling high-dimensional or complex data in Bayesian filtering problems. This innovative approach incorporates importance weights and an iterative re-weighting scheme inspired by iteratively re-weighted Least Squares (IRLS) to enhance the robustness and accuracy of Bayesian inference. The proposed method does not require explicit specification of prior and likelihood distributions; instead, it learns the kernel mean representations from training data. Experimental results demonstrate the superior performance of this method over traditional KBF methods on high-dimensional datasets.

High‐temperature ablation resistance prediction of ceramic coatings using machine learning

AbstractSurface ablation temperature and linear ablation rate are two crucial indicators for ceramic coatings under ultrahigh temperatures service, yet the results collection of such two indicators in the process is difficult due to the long‐period material preparation and the high‐cost test. In this work, four kinds of machine learning models are applied to predict the above two indicators. The Random Forest (RF) model exhibits a high accuracy of 87% in predicting surface ablation temperature, while a low accuracy of 60% in linear ablation rate. To optimize the model, the novel features are constructed based on the original features by the sum of the importance weights in the model. Thereafter, the importance of the newly constructed features increases significantly, and the accuracy of the optimized RF model is improved by 11%, exceeding 70% in accuracy. By validation with available data and experiments, the optimized model demonstrates precise predictions of the target variables.

An Operational Quality Model of Embedded Software Aligned with ISO 25000

Embedded systems omnipresent in everyday life and industry are mainly composed of hardware and software that must comply with a number of standards and regulations. However, there is no consensus on the quality characteristics and sub-characteristics of embedded software. This paper presents the steps for modeling an operational quality model for embedded software aligned with the ISO 25000 series of quality models for traditional computer systems. From a literature review composed of 40 studies on quality modeling for embedded systems and software, 85 of the most frequent quality characteristics and sub-characteristics were first identified, including a subset of 16 referenced or cited in at least 25% of the literature. Next, the design of a quality model for embedded software aligned with the ISO 25000 series was proposed with 13 characteristics and 27 sub-characteristics. The operational aspect of this quality model for embedded software is addressed next through a set of measures and measurement functions from ISO 25000 to aggregate the results of the quantification of the characteristics and sub-characteristics. A survey involving 25 embedded software specialists is presented next to gauge, using Fleiss’ Kappa criteria, their agreement with the proposed quality model. Furthermore, the computed importance weights derived from the survey participants’ individual opinions were compared with those derived from an analysis of 40 embedded software studies, bolstering the credibility of the model. The results of this study suggest that the proposed quality model can serve as a framework for evaluating and understanding the quality characteristics across diverse expertise levels. Furthermore, the convergence between the survey and the literature strengthens the model's credibility by anchoring it in both established literature and practitioners’ agreements.

Performance assessment of agricultural waste based eco‐friendly brake friction composites

AbstractThere has been growing interest in using sustainable and eco‐friendly products to produce engineering materials. For this purpose, composite material applications obtained from agricultural wastes are gaining popularity. This study examines the synergistic effect of rice husk and rice stalk wastes on the fade and recovery performance of brake friction composites. Brake friction materials were developed using rice husk and rice stalk separately and in two different weight percentages as a 5–10 ratio in the formulation. For comparison purposes, a reference brake pad using copper as a substitute and a commercially available brake pad were used. Various physical, mechanical and thermal properties were analyzed. The tribological behavior of friction composites was evaluated on the Krauss test device in line with the ECE R90 procedure. The worn surface properties were analyzed using scanning electron microscopy. Tribo test results of friction composites were taken as criteria for performance optimization. While the importance weight of the criteria was determined by AHP, the VIKOR method was used in the sorting of alternatives. The experimental results have revealed that rice husk‐added friction composites had a good coefficient of friction value with better fade and recovery performance compared to rice stalk‐added ones. Increasing the amount of rice husk and rice stalk in the formulation tended to decrease the fade performance; however, it has increased the wear rate and recovery properties. Optimization results have shown that the brake friction composite containing 5 wt % rice husk ranks first in meeting the desired tribological criteria.Highlights Cu‐free rice husk and rice stalk‐added friction composites were developed. Fibrous structure in the matrix developed the contact plateaus. Rice husk‐based tribo‐layer protected the composite from further wear damage. The addition of agro‐waste to friction composites exhibited good potential.