Pub Date : 2021-08-17DOI: 10.1080/00224065.2021.1960936
Wendong Li, F. Tsung, Zhenli Song, Ke Zhang, D. Xiang
Abstract Landslides are severe geographical activities that result in large quantities of rock and debris flowing down hill-slopes, leading to thousands of casualties and billions of dollars in infrastructure damage every year worldwide. For detecting landslides, on-site sensor systems are widely applied for data collection and many existing statistical process control methods can be adopted for modeling and monitoring. However, the conventional methods may perform poorly or even inapplicable when the sensors have different set-up times and end times, especially when the system includes newly deployed sensors with limited data collected. To make effective use of such new sensors immediately after deployment, we propose a novel multi-sensor based charting scheme for dynamic landslide modeling and monitoring by using transfer learning. A regularized parameter-based transfer learning approach integrated with the ordered LASSO is first proposed to effectively transfer information from old sensors with sufficient historical data to new ones with limited data. The approach considers the similarities not only between the autoregressive coefficients of different sensors, but also between the temporal correlation patterns. A control chart is then proposed for monitoring the newly deployed sensors sequentially based on the generalized likelihood ratio. Extensive simulation results and a real data example of landslide monitoring demonstrate the effectiveness of our proposed method.
{"title":"Multi-sensor based landslide monitoring via transfer learning","authors":"Wendong Li, F. Tsung, Zhenli Song, Ke Zhang, D. Xiang","doi":"10.1080/00224065.2021.1960936","DOIUrl":"https://doi.org/10.1080/00224065.2021.1960936","url":null,"abstract":"Abstract Landslides are severe geographical activities that result in large quantities of rock and debris flowing down hill-slopes, leading to thousands of casualties and billions of dollars in infrastructure damage every year worldwide. For detecting landslides, on-site sensor systems are widely applied for data collection and many existing statistical process control methods can be adopted for modeling and monitoring. However, the conventional methods may perform poorly or even inapplicable when the sensors have different set-up times and end times, especially when the system includes newly deployed sensors with limited data collected. To make effective use of such new sensors immediately after deployment, we propose a novel multi-sensor based charting scheme for dynamic landslide modeling and monitoring by using transfer learning. A regularized parameter-based transfer learning approach integrated with the ordered LASSO is first proposed to effectively transfer information from old sensors with sufficient historical data to new ones with limited data. The approach considers the similarities not only between the autoregressive coefficients of different sensors, but also between the temporal correlation patterns. A control chart is then proposed for monitoring the newly deployed sensors sequentially based on the generalized likelihood ratio. Extensive simulation results and a real data example of landslide monitoring demonstrate the effectiveness of our proposed method.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82028594","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-08-17DOI: 10.1080/00224065.2021.1960934
Minhee Kim, Jing-Ru C. Cheng, Kaibo Liu
Abstract Recent advances in sensor technology have made it possible to monitor the degradation of a system using multiple sensors simultaneously. Accordingly, many neural network-based prognostic models have been proposed to use observed multiple sensor signals as inputs and estimate the degradation status or failure time of the system. Although these models have achieved promising prognostic performance, it is still difficult to interpret the extracted features, and the models are often used in a black-box manner providing only the final results. In this study, a novel sensor selection framework is proposed to address this challenge by adaptively deciding which sensors to use at the moment to enhance remaining useful life prediction. The contributions of this work are summarized as follows: (1) being generic and can be attached to a variety of existing neural network-based prognostic models; (2) being trained in a unified manner to optimize both the sensor selection and prognostic accuracies simultaneously; (3) improving the interpretability of the model by explaining how different sensors contribute to the final remaining useful life prediction of individual systems over time; and (4) introducing several regularization techniques to ensure the stability of the training process. We validate the proposed framework using a series of numerical studies on the degradation of aircraft gas turbine engines.
{"title":"An adaptive sensor selection framework for multisensor prognostics","authors":"Minhee Kim, Jing-Ru C. Cheng, Kaibo Liu","doi":"10.1080/00224065.2021.1960934","DOIUrl":"https://doi.org/10.1080/00224065.2021.1960934","url":null,"abstract":"Abstract Recent advances in sensor technology have made it possible to monitor the degradation of a system using multiple sensors simultaneously. Accordingly, many neural network-based prognostic models have been proposed to use observed multiple sensor signals as inputs and estimate the degradation status or failure time of the system. Although these models have achieved promising prognostic performance, it is still difficult to interpret the extracted features, and the models are often used in a black-box manner providing only the final results. In this study, a novel sensor selection framework is proposed to address this challenge by adaptively deciding which sensors to use at the moment to enhance remaining useful life prediction. The contributions of this work are summarized as follows: (1) being generic and can be attached to a variety of existing neural network-based prognostic models; (2) being trained in a unified manner to optimize both the sensor selection and prognostic accuracies simultaneously; (3) improving the interpretability of the model by explaining how different sensors contribute to the final remaining useful life prediction of individual systems over time; and (4) introducing several regularization techniques to ensure the stability of the training process. We validate the proposed framework using a series of numerical studies on the degradation of aircraft gas turbine engines.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89722065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-08-12DOI: 10.1080/00224065.2021.1960935
Nicholas Rios, Dennis K. J. Lin
Abstract In a mixture experiment, m components are mixed to produce a response. The total amount of the mixture is a constant. Existing literature on mixture designs ignores the order of addition of the mixture components. This paper considers the Order-of-Addition (OofA) mixture experiment, where the response depends on both the mixture proportions of components and their order of addition. Empirical study demonstrates that if mixture-order interactions exist, then the optimal mixture proportions identified by traditional models may be misleading. Full Mixture OofA designs are created which ensure orthogonality between mixture model terms and addition order effects. These designs allow for the estimation of (1) typical mixture model parameters and (2) order-of-addition effects. Moreover, models which include both main effects and key mixture-order interactions are introduced.
{"title":"Order-of-addition mixture experiments","authors":"Nicholas Rios, Dennis K. J. Lin","doi":"10.1080/00224065.2021.1960935","DOIUrl":"https://doi.org/10.1080/00224065.2021.1960935","url":null,"abstract":"Abstract In a mixture experiment, m components are mixed to produce a response. The total amount of the mixture is a constant. Existing literature on mixture designs ignores the order of addition of the mixture components. This paper considers the Order-of-Addition (OofA) mixture experiment, where the response depends on both the mixture proportions of components and their order of addition. Empirical study demonstrates that if mixture-order interactions exist, then the optimal mixture proportions identified by traditional models may be misleading. Full Mixture OofA designs are created which ensure orthogonality between mixture model terms and addition order effects. These designs allow for the estimation of (1) typical mixture model parameters and (2) order-of-addition effects. Moreover, models which include both main effects and key mixture-order interactions are introduced.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88300537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-08-05DOI: 10.1080/00224065.2021.1948372
Qian Wang, K. Paynabar, M. Pacella
Abstract Faults occurred during the operational lifetime of photovoltaic (PV) systems can cause energy loss, system shutdown, as well as possible fire risks. Therefore, it is crucial to detect anomalies and faults to control the system’s performance and ensure its reliability. Comparing to traditional monitoring techniques based on an on-site visual inspection and/ or electrical measuring devices, the combination of drones and infrared thermography imaging evidently provides the means for faster and less expensive PV monitoring. However, the literature in this area lacks automatic and implementable algorithms for PV fault detection, particularly, using raw aerial thermography, with precise performance evaluation. The objective of this paper is, thus, to build a fully automatic online monitoring framework. We propose an analytical framework for online analysis of the raw video streams of aerial thermography. This framework integrates image processing and statistical machine learning techniques. We validate the effectiveness of the proposed framework and provide sufficient details to facilitate its implementation by practitioners. Two challenges hinder direct fault detection on raw PV images. One is that raw PV images often have non-smooth backgrounds that can impact the detection performance. This background needs to be removed before fault detection. However, this is a daunting task given the perspective of images. To deal with this challenge, we utilize the Transform Invariant Low-rank Textures (TILT) method to orthogonalize the perspective before applying edge detection to crop out the background and aligning the cropped images. The other issue is that the regular hot spots at the bottom edges of the solar panels are normal and should not be detected as anomalies. This makes the intensity-based detection method in the literature fail. These hot spots are part of the low-rank pattern of the image sequence. On the other hand, the hot spots caused by anomalies deviate from the normal low-rank pattern of the PV cells. Therefore, we propose a methodology that relies on Robust Principal Component Analysis (RPCA), which can separate sparse corrupted anomalous components from a low-rank background. The RPCA is applied to the PV images for simultaneous detection and isolation of anomalies. In addition to RPCA, we suggest a set of post-processing procedures for image denoising, and segmentation. The proposed algorithm is developed using 20 normal (with no anomalies) training samples and 100 test samples. The results showed that the algorithm successfully detects the anomalies with a recall of 0.80 and detects the significant anomalies with the maximum recall of 1. Our method outperforms two benchmark methods in terms of F1 score by 44.5% and 114.3%. The small number of false alarms is mostly due to irregular image patterns at the end of a PV array or an extreme non-orthogonal perspective. Since the number of false alarms is not large, it does not disrupt
{"title":"Online automatic anomaly detection for photovoltaic systems using thermography imaging and low rank matrix decomposition","authors":"Qian Wang, K. Paynabar, M. Pacella","doi":"10.1080/00224065.2021.1948372","DOIUrl":"https://doi.org/10.1080/00224065.2021.1948372","url":null,"abstract":"Abstract Faults occurred during the operational lifetime of photovoltaic (PV) systems can cause energy loss, system shutdown, as well as possible fire risks. Therefore, it is crucial to detect anomalies and faults to control the system’s performance and ensure its reliability. Comparing to traditional monitoring techniques based on an on-site visual inspection and/ or electrical measuring devices, the combination of drones and infrared thermography imaging evidently provides the means for faster and less expensive PV monitoring. However, the literature in this area lacks automatic and implementable algorithms for PV fault detection, particularly, using raw aerial thermography, with precise performance evaluation. The objective of this paper is, thus, to build a fully automatic online monitoring framework. We propose an analytical framework for online analysis of the raw video streams of aerial thermography. This framework integrates image processing and statistical machine learning techniques. We validate the effectiveness of the proposed framework and provide sufficient details to facilitate its implementation by practitioners. Two challenges hinder direct fault detection on raw PV images. One is that raw PV images often have non-smooth backgrounds that can impact the detection performance. This background needs to be removed before fault detection. However, this is a daunting task given the perspective of images. To deal with this challenge, we utilize the Transform Invariant Low-rank Textures (TILT) method to orthogonalize the perspective before applying edge detection to crop out the background and aligning the cropped images. The other issue is that the regular hot spots at the bottom edges of the solar panels are normal and should not be detected as anomalies. This makes the intensity-based detection method in the literature fail. These hot spots are part of the low-rank pattern of the image sequence. On the other hand, the hot spots caused by anomalies deviate from the normal low-rank pattern of the PV cells. Therefore, we propose a methodology that relies on Robust Principal Component Analysis (RPCA), which can separate sparse corrupted anomalous components from a low-rank background. The RPCA is applied to the PV images for simultaneous detection and isolation of anomalies. In addition to RPCA, we suggest a set of post-processing procedures for image denoising, and segmentation. The proposed algorithm is developed using 20 normal (with no anomalies) training samples and 100 test samples. The results showed that the algorithm successfully detects the anomalies with a recall of 0.80 and detects the significant anomalies with the maximum recall of 1. Our method outperforms two benchmark methods in terms of F1 score by 44.5% and 114.3%. The small number of false alarms is mostly due to irregular image patterns at the end of a PV array or an extreme non-orthogonal perspective. Since the number of false alarms is not large, it does not disrupt","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78111308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-30DOI: 10.1080/00224065.2021.1951147
A. Iquebal
{"title":"Introduction to time series modeling with applications in R","authors":"A. Iquebal","doi":"10.1080/00224065.2021.1951147","DOIUrl":"https://doi.org/10.1080/00224065.2021.1951147","url":null,"abstract":"","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85825395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-27DOI: 10.1080/00224065.2021.1958720
Caleb King
Presenting a thorough review of selected topics in survival analysis, Dr. Legrand’s Advanced Survival Models is an excellent reference for students and/or practitioners. The book covers four advanced topics: frailty models, cure models, competing risks, and joint modeling of time-dependent covariates. Each topic is addressed with great care, balancing coverage with intimate detail so that readers come away with a comfortable level of knowledge about each topic. The book is divided into six chapters. The first chapter covers basic survival analysis concepts and presents six medical datasets (many of which are publicly available) for illustrating the models going forward. Each dataset is explained in great detail, including the context of data collection and the meaning of each variable. The data are analyzed using primarily R code throughout with a sprinkling of SAS code as well. The second chapter then presents a brief review of classical survival analysis techniques. It is in this chapter that readers get a taste of the level of detail with which Dr. Legrand discusses each of the advanced models: parametric models, semi-parametric models, non-parametric models, Cox proportional hazards, accelerated failure time; all are given their due diligence and illustrated with the data provided, so that the reader is presented with the breadth of methodology available at even the basic level. For the remaining four chapters, the format is similar. The chapter opens with an overall introduction of the topic, effectively summarizing the contents to come. Next, the primary model varieties are presented with sufficient context to understand their origins as well as their areas of appropriate use. Next, the primary methods of estimating the models are discussed. Finally, the chapter ends with illustration of the models using one or more of the datasets. In each case, Dr. Legrand presents enough detail so that the reader becomes intimately familiar with the basic concepts and estimation procedures. As an illustration of her effectiveness, I was not aware of the existence of cure models prior to reading this book. Now, I feel confident enough on the subject that I would be comfortable explaining it to another person. Where there is the opportunity for more specialized models and/or estimation procedures, multiple references are provided that discuss such models and/or procedures in greater detail. I found the references satisfactory for further study on a particular topic. While the book overall is a fairly easy read, there are several editorial “glitches” that, though not sufficient to cause confusion and misunderstanding, were still noticeable and tended to happen more frequently than one would expect. Most of these “glitches” consist of typographical errors and awkward sentence structures. In addition, the material was a bit repetitive when moving from the introduction to the main material in each chapter. However, I consider both of these to be very minor inc
{"title":"Advanced Survival Models","authors":"Caleb King","doi":"10.1080/00224065.2021.1958720","DOIUrl":"https://doi.org/10.1080/00224065.2021.1958720","url":null,"abstract":"Presenting a thorough review of selected topics in survival analysis, Dr. Legrand’s Advanced Survival Models is an excellent reference for students and/or practitioners. The book covers four advanced topics: frailty models, cure models, competing risks, and joint modeling of time-dependent covariates. Each topic is addressed with great care, balancing coverage with intimate detail so that readers come away with a comfortable level of knowledge about each topic. The book is divided into six chapters. The first chapter covers basic survival analysis concepts and presents six medical datasets (many of which are publicly available) for illustrating the models going forward. Each dataset is explained in great detail, including the context of data collection and the meaning of each variable. The data are analyzed using primarily R code throughout with a sprinkling of SAS code as well. The second chapter then presents a brief review of classical survival analysis techniques. It is in this chapter that readers get a taste of the level of detail with which Dr. Legrand discusses each of the advanced models: parametric models, semi-parametric models, non-parametric models, Cox proportional hazards, accelerated failure time; all are given their due diligence and illustrated with the data provided, so that the reader is presented with the breadth of methodology available at even the basic level. For the remaining four chapters, the format is similar. The chapter opens with an overall introduction of the topic, effectively summarizing the contents to come. Next, the primary model varieties are presented with sufficient context to understand their origins as well as their areas of appropriate use. Next, the primary methods of estimating the models are discussed. Finally, the chapter ends with illustration of the models using one or more of the datasets. In each case, Dr. Legrand presents enough detail so that the reader becomes intimately familiar with the basic concepts and estimation procedures. As an illustration of her effectiveness, I was not aware of the existence of cure models prior to reading this book. Now, I feel confident enough on the subject that I would be comfortable explaining it to another person. Where there is the opportunity for more specialized models and/or estimation procedures, multiple references are provided that discuss such models and/or procedures in greater detail. I found the references satisfactory for further study on a particular topic. While the book overall is a fairly easy read, there are several editorial “glitches” that, though not sufficient to cause confusion and misunderstanding, were still noticeable and tended to happen more frequently than one would expect. Most of these “glitches” consist of typographical errors and awkward sentence structures. In addition, the material was a bit repetitive when moving from the introduction to the main material in each chapter. However, I consider both of these to be very minor inc","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78408105","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-21DOI: 10.1080/00224065.2021.1947162
Burcu Aytaçoğlu, Anne R. Driscoll, W. Woodall
Abstract An integral part of the design of control charts, including the multivariate exponentially weighted moving average (MEWMA) control chart, is the determination of the appropriate control limits for prospective monitoring. Methods using Markov chain analyses, integral equations, and simulation have been proposed to determine the MEWMA chart limits when the limits are based on a specified in-control average run length (ARL) value. A drawback of the usual approach is that the conditional false alarm rate (CFAR) for these charts varies over time in what might be in an unexpected and undesirable way. We define the CFAR as the probability of a false alarm given no previous false alarm. We do not condition on the results of a Phase I sample, as done by others, in studies of the effect of estimation error on control chart performance. We propose the use of dynamic probability control limits (DPCLs) to keep the CFAR constant over time at a specified value. The CFAR at any time, however, could be controlled to be any specified value using our approach. Using simulation, we determine the DPCLs for the MEWMA control chart being used to monitor the mean vector with an assumed known variance-covariance matrix. We consider cases where the sample size is both fixed and time-varying. For varying sample sizes, the DPCLs adapt automatically to any change in the sample size distribution. In all cases, the CFAR is held closely to a fixed value and the resulting in-control run length performance follows closely to that of the geometric distribution.
{"title":"Controlling the conditional false alarm rate for the MEWMA control chart","authors":"Burcu Aytaçoğlu, Anne R. Driscoll, W. Woodall","doi":"10.1080/00224065.2021.1947162","DOIUrl":"https://doi.org/10.1080/00224065.2021.1947162","url":null,"abstract":"Abstract An integral part of the design of control charts, including the multivariate exponentially weighted moving average (MEWMA) control chart, is the determination of the appropriate control limits for prospective monitoring. Methods using Markov chain analyses, integral equations, and simulation have been proposed to determine the MEWMA chart limits when the limits are based on a specified in-control average run length (ARL) value. A drawback of the usual approach is that the conditional false alarm rate (CFAR) for these charts varies over time in what might be in an unexpected and undesirable way. We define the CFAR as the probability of a false alarm given no previous false alarm. We do not condition on the results of a Phase I sample, as done by others, in studies of the effect of estimation error on control chart performance. We propose the use of dynamic probability control limits (DPCLs) to keep the CFAR constant over time at a specified value. The CFAR at any time, however, could be controlled to be any specified value using our approach. Using simulation, we determine the DPCLs for the MEWMA control chart being used to monitor the mean vector with an assumed known variance-covariance matrix. We consider cases where the sample size is both fixed and time-varying. For varying sample sizes, the DPCLs adapt automatically to any change in the sample size distribution. In all cases, the CFAR is held closely to a fixed value and the resulting in-control run length performance follows closely to that of the geometric distribution.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79977580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-15DOI: 10.1080/00224065.2021.1939815
Miao Cai, Amir Mehdizadeh, Qiong Hu, Mohammad Ali Alamdar Yazdi, A. Vinel, Karen C. Davis, Hong Xian, F. Megahed, S. Rigdon
Abstract Quality in the trucking industry involves several facets, including on-time performance and safety. In the largest naturalistic driving study to-date, with 496 drivers and 13 M miles driven, we address two safety questions: (a) does the occurrence of safety critical events increase during a driving shift? and (b) what is the effect of rest breaks on the incidence of those events? To address these two questions, we adopt point process models, commonly used to assess the reliability of repairable systems, to model the occurrence/likelihood of safety critical events. To account for driver differences, driver-level random effects were also assumed. Our results show that: (a) the intensity for hard brakes decreases throughout a shift, (b) rest breaks reduce the likelihood of activation of the automated collision mitigation system, and (c) there is a fair amount of variability among drivers. Given that a hard brake (a less severe safety critical event) is more common in the beginning of the shift, it can potentially be explained by an increased likelihood of being in a local/city road and/or increased likelihood of aggressive driving behavior early in a driver’s shift. Furthermore, we quantified the impact of rest breaks in reducing engagement of the more severe automated collision mitigation system, providing data-driven evidence on the importance of rest-break scheduling for trucking safety. Properties of the approach were also investigated through a simulation study, where we examined the consequences of an incorrect specification of the Bayesian priors.
{"title":"Hierarchical point process models for recurring safety critical events involving commercial truck drivers: A reliability framework for human performance modeling","authors":"Miao Cai, Amir Mehdizadeh, Qiong Hu, Mohammad Ali Alamdar Yazdi, A. Vinel, Karen C. Davis, Hong Xian, F. Megahed, S. Rigdon","doi":"10.1080/00224065.2021.1939815","DOIUrl":"https://doi.org/10.1080/00224065.2021.1939815","url":null,"abstract":"Abstract Quality in the trucking industry involves several facets, including on-time performance and safety. In the largest naturalistic driving study to-date, with 496 drivers and 13 M miles driven, we address two safety questions: (a) does the occurrence of safety critical events increase during a driving shift? and (b) what is the effect of rest breaks on the incidence of those events? To address these two questions, we adopt point process models, commonly used to assess the reliability of repairable systems, to model the occurrence/likelihood of safety critical events. To account for driver differences, driver-level random effects were also assumed. Our results show that: (a) the intensity for hard brakes decreases throughout a shift, (b) rest breaks reduce the likelihood of activation of the automated collision mitigation system, and (c) there is a fair amount of variability among drivers. Given that a hard brake (a less severe safety critical event) is more common in the beginning of the shift, it can potentially be explained by an increased likelihood of being in a local/city road and/or increased likelihood of aggressive driving behavior early in a driver’s shift. Furthermore, we quantified the impact of rest breaks in reducing engagement of the more severe automated collision mitigation system, providing data-driven evidence on the importance of rest-break scheduling for trucking safety. Properties of the approach were also investigated through a simulation study, where we examined the consequences of an incorrect specification of the Bayesian priors.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81797217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-12DOI: 10.1080/00224065.2021.1947163
Shuai Huang
This book offers a good introduction to some statistical methods used in elections. It has two parts. The first part contains four chapters that cover estimation methods for polls. The main technical problem is the estimation of the proportion of the population holding a particular preference in voting. The analytic core of the problem is binomial distribution and the sampling and estimation procedures center around this distribution. As in the real world there are always complications. Various remedies are provided to address these complications. The author has done a great job of introducing the critical concepts and considerations in both the problem formulation and solution. For example, when introducing the importance of weighting in deriving the estimate of the poll, the author pretends to write a press release of his poll result on the issue of gender fairness in the military. It is clear that if a different source of demographics statistics is used, the poll result is quite different. Examples as such are quite useful for readers to understand the subject matter and its complexity. The second part of the book covers a few techniques to detect frauds and anomalies by examining the election results. Some techniques build on an interesting premise that humans are bad at mimicking randomness. This echoes what Fisher (1958) had said, “if one tries to think of numbers at random, one thinks of numbers very far from at random.” The Benford test is introduced in detail, including its history and its interesting applications in analyzing election data to detect anomaly based on the distributions of the leading digits reported by different precincts. The differential invalidation and some regression models are introduced as well. Spatial correlations could be modeled by using the geographical information in the data. The book concludes with a detailed discussion on data from Sri Lanka since 1994. This is a useful book that can help a broad range of readers to appreciate the power of statistics in understanding the election process from an analytic and scientific perspective. On top of the techniques introduced in the book, there are anecdotes and comments and insights that can enrich the reading experience. E.g., as in the preface the statement from a Nicaraguan leader “Indeed, you won the elections, but I won the count.” or the comment in the end of Chapter 4 “as with many things in statistics, increasing quality in one area tends to reduce quality in another.” Statistical techniques in this book are tightly bonded with the contexts and the backgrounds of their application. After reading the book, I appreciate the book has helped me understand a complex problem in a complex world. Not everything is what it appears to be, but we can equip ourselves with sufficient knowledge and useful tools to help us look at the data in every angle and really feel the data as it is.
{"title":"Understanding elections through statistics by Ole J. Forsberg, CRC press, Taylor & Francis group, boca Raton, FL, 2020, 225 pp., $69.95, ISBN 978-0367895372","authors":"Shuai Huang","doi":"10.1080/00224065.2021.1947163","DOIUrl":"https://doi.org/10.1080/00224065.2021.1947163","url":null,"abstract":"This book offers a good introduction to some statistical methods used in elections. It has two parts. The first part contains four chapters that cover estimation methods for polls. The main technical problem is the estimation of the proportion of the population holding a particular preference in voting. The analytic core of the problem is binomial distribution and the sampling and estimation procedures center around this distribution. As in the real world there are always complications. Various remedies are provided to address these complications. The author has done a great job of introducing the critical concepts and considerations in both the problem formulation and solution. For example, when introducing the importance of weighting in deriving the estimate of the poll, the author pretends to write a press release of his poll result on the issue of gender fairness in the military. It is clear that if a different source of demographics statistics is used, the poll result is quite different. Examples as such are quite useful for readers to understand the subject matter and its complexity. The second part of the book covers a few techniques to detect frauds and anomalies by examining the election results. Some techniques build on an interesting premise that humans are bad at mimicking randomness. This echoes what Fisher (1958) had said, “if one tries to think of numbers at random, one thinks of numbers very far from at random.” The Benford test is introduced in detail, including its history and its interesting applications in analyzing election data to detect anomaly based on the distributions of the leading digits reported by different precincts. The differential invalidation and some regression models are introduced as well. Spatial correlations could be modeled by using the geographical information in the data. The book concludes with a detailed discussion on data from Sri Lanka since 1994. This is a useful book that can help a broad range of readers to appreciate the power of statistics in understanding the election process from an analytic and scientific perspective. On top of the techniques introduced in the book, there are anecdotes and comments and insights that can enrich the reading experience. E.g., as in the preface the statement from a Nicaraguan leader “Indeed, you won the elections, but I won the count.” or the comment in the end of Chapter 4 “as with many things in statistics, increasing quality in one area tends to reduce quality in another.” Statistical techniques in this book are tightly bonded with the contexts and the backgrounds of their application. After reading the book, I appreciate the book has helped me understand a complex problem in a complex world. Not everything is what it appears to be, but we can equip ourselves with sufficient knowledge and useful tools to help us look at the data in every angle and really feel the data as it is.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80102706","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-03DOI: 10.1080/00224065.2021.1948373
Xiao Liu, Rong Pan
Abstract Recurrence data arise from multi-disciplinary domains spanning reliability, cyber security, healthcare, online retailing, etc. This paper investigates an additive-tree-based approach, known as Boost-R (Boosting for Recurrence Data), for recurrent event data with both static and dynamic features. Boost-R constructs an ensemble of gradient boosted additive trees to estimate the cumulative intensity function of the recurrent event process, where a new tree is added to the ensemble by minimizing the regularized L 2 distance between the observed and predicted cumulative intensity. Unlike conventional regression trees, a time-dependent function is constructed by Boost-R on each tree leaf. The sum of these functions, from multiple trees, yields the ensemble estimator of the cumulative intensity. The divide-and-conquer nature of tree-based methods is appealing when hidden sub-populations exist within a heterogeneous population. The non-parametric nature of regression trees helps to avoid parametric assumptions on the complex interactions between event processes and features. Critical insights and advantages of Boost-R are investigated through comprehensive numerical examples. Datasets and computer code of Boost-R are made available on GitHub. To our best knowledge, Boost-R is the first gradient boosted additive-tree-based approach for modeling large-scale recurrent event data with both static and dynamic feature information.
重复数据来自多学科领域，包括可靠性、网络安全、医疗保健、在线零售等。本文研究了一种基于加性树的方法，称为Boost-R (Boosting for recurrent Data)，用于具有静态和动态特征的循环事件数据。Boost-R构建了一个梯度增强的加性树集合来估计循环事件过程的累积强度函数，其中通过最小化观测到的和预测的累积强度之间的正则化l2距离，将新树添加到集合中。与传统的回归树不同，Boost-R在每个树叶上构建了一个时间相关的函数。这些函数的和，从多个树，产生累积强度的集合估计。当隐藏的子种群存在于异质种群中时，基于树的方法的分而治之的特性很有吸引力。回归树的非参数性质有助于避免对事件过程和特征之间复杂的相互作用进行参数假设。通过全面的数值实例研究了Boost-R的关键见解和优势。Boost-R的数据集和计算机代码可在GitHub上获得。据我们所知，Boost-R是第一个基于梯度增强加性树的方法，用于对具有静态和动态特征信息的大规模循环事件数据建模。
{"title":"Boost-R: Gradient boosted trees for recurrence data","authors":"Xiao Liu, Rong Pan","doi":"10.1080/00224065.2021.1948373","DOIUrl":"https://doi.org/10.1080/00224065.2021.1948373","url":null,"abstract":"Abstract Recurrence data arise from multi-disciplinary domains spanning reliability, cyber security, healthcare, online retailing, etc. This paper investigates an additive-tree-based approach, known as Boost-R (Boosting for Recurrence Data), for recurrent event data with both static and dynamic features. Boost-R constructs an ensemble of gradient boosted additive trees to estimate the cumulative intensity function of the recurrent event process, where a new tree is added to the ensemble by minimizing the regularized L 2 distance between the observed and predicted cumulative intensity. Unlike conventional regression trees, a time-dependent function is constructed by Boost-R on each tree leaf. The sum of these functions, from multiple trees, yields the ensemble estimator of the cumulative intensity. The divide-and-conquer nature of tree-based methods is appealing when hidden sub-populations exist within a heterogeneous population. The non-parametric nature of regression trees helps to avoid parametric assumptions on the complex interactions between event processes and features. Critical insights and advantages of Boost-R are investigated through comprehensive numerical examples. Datasets and computer code of Boost-R are made available on GitHub. To our best knowledge, Boost-R is the first gradient boosted additive-tree-based approach for modeling large-scale recurrent event data with both static and dynamic feature information.","PeriodicalId":54769,"journal":{"name":"Journal of Quality Technology","volume":null,"pages":null},"PeriodicalIF":2.5,"publicationDate":"2021-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78715817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：