Adel Alaeddini

ASRSM: A Sequential Experimental Design for Response Surface Optimization

Most preset response surface methodology (RSM) designs offer ease of implementation and good performance over a wide range of process and design optimization applications. These designs often lack the ability to adapt the design on the basis of the characteristics of application and experimental space so as to reduce the number of experiments necessary. Hence, they are not cost-effective for applications where the cost of experimentation is high or when the experimentation resources are limited. In this paper, we present an adaptive sequential response surface methodology (ASRSM) for industrial experiments with high experimentation cost, limited experimental resources, and high design optimization performance requirement. The proposed approach is a sequential adaptive experimentation approach that combines concepts from nonlinear optimization, design of experiments, and response surface optimization. The ASRSM uses the information gained from the previous experiments to design the subsequent experiment by simultaneously reducing the region of interest and identifying factor combinations for new experiments. Its major advantage is the experimentation efficiency such that for a given response target, it identifies the input factor combination (or containing region) in less number of experiments than the classical single-shot RSM designs. Through extensive simulated experiments and real-world case studies, we show that the proposed ASRSM method outperforms the popular central composite design method and compares favorably with optimal designs. Copyright

A hybrid prediction model for no-shows and cancellations of outpatient appointments

A no-show occurs when a scheduled patient neither keeps nor cancels the appointment. A cancellation happens when individuals contact the clinic and cancel their scheduled appointments. Such disruptions not only cause inconvenience to hospital management, they also have a significant impact on the revenue, cost and resource utilization for almost all of the healthcare systems. In this paper, we develop a hybrid probabilistic model based on multinomial logistic regression and Bayesian inference to predict accurately the probability of no-show and cancellation in real-time. First, a multinomial logistic regression model is built based on the entire populations general social and demographic information to provide initial estimates of no-show and cancellation probabilities. Next, the estimated probabilities from the logistic model are transformed into a bivariate Dirichlet distribution, which is used as the prior distribution of a Bayesian updating mechanism to personalize the initial estimates for each patient based on his/her attendance record. In addition, to further improve the estimates, prior to applying the Bayesian updating mechanism, each appointment in the database is weighted based on its recency, weekday of occurrence, and clinic type. The effectiveness of the proposed approach is demonstrated using healthcare data collected at a medical center. We also discuss the advantages of the proposed hybrid model and describe possible real-world applications.

An Efficient Adaptive Sequential Methodology for Expensive Response Surface Optimization

The preset response surface methodology (RSM) designs are commonly used in a wide range of process and design optimization applications. Although they offer ease of implementation and good performance, they are not sufficiently adaptive to reduce the required number of experiments and thus are not cost effective for applications with high cost of experimentation. We propose an efficient adaptive sequential methodology based on optimal design and experiments ranking for response surface optimization (O-ASRSM) for industrial experiments with high experimentation cost, limited experimental resources, and requiring high design optimization performance. The proposed approach combines the concepts from optimal design of experiments, nonlinear optimization, and RSM. By using the information gained from the previous experiments, O-ASRSM designs the subsequent experiment by simultaneously reducing the region of interest and by identifying factor combinations for new experiments. Given a given response target, O-ASRSM identifies the input factor combination in less number of experiments than the classical single-shot RSM designs. We conducted extensive simulated experiments involving quadratic and nonlinear response functions. The results show that the O-ASRSM method outperforms the popular central composite design, the Box–Behnken design, and the optimal designs and is competitive with other sequential response surface methods in the literature. Furthermore, results indicate that O-ASRSM’s performance is robust with respect to the increasing number of factors. Copyright

Multivariate adaptive approach for monitoring simple linear profiles

Adaptive sample size and sampling intervals schemes have been widely used to improve the statistical efficiency of Hotelling T² control chart in detecting small changes when the quality of a product or a process can be characterised by the multivariate distribution of quality characteristics. In this paper, we design a Hotelling T² scheme varying sample sizes and sampling intervals VSSI-T² for accelerating the speed of detecting off-target conditions in linear profile parameters. We investigate the statistical performance of the adaptive approach versus its fixed sampling counterparts. To find the optimal setting of the VSSI-T², we build an optimisation model solved using genetic algorithm GA. Also, average time to signal ATS is considered as the objective function of the model and estimated using the Markov chain fundamentals. The comparative studies reveal the potentials of the adaptive scheme in improving the performance of the Hotelling T² control chart in monitoring linear profiles.

An adaptive sequential experimentation methodology for expensive response surface optimization - Case study in traumatic brain injury modeling

The preset response surface designs often lack the ability to adapt the design based on the characteristics of application and experimental space so as to reduce the number of experiments necessary. Hence, they are not cost effective for applications where the cost of experimentation is high or when the experimentation resources are limited. In this paper, we present an adaptive sequential methodology for n-dimensional response surface optimization (n-dimensional adaptive sequential response surface methodology (N-ASRSM)) for industrial experiments with high experimentation cost, which requires high design optimization performance. We also develop a novel risk adjustment strategy for effectively considering the effect of noise into the design. The N-ASRSM is a sequential adaptive experimentation approach, which uses the information from previous experiments to design the subsequent experiment by simultaneously reducing the region of interest and identifying factor combinations for new experiments. Its major advantage is the experimentation efficiency such that, for a given response target, it identifies the input factor combination in less number of experiments than the classical response surface methodology designs. We applied N-ASRSM to the problem of traumatic brain injury modeling and compared the result with the conventional central composite design. Also, through extensive simulated experiments with different quadratic and nonlinear cases, we show that the proposed N-ASRSM method outperforms the classical response surface methodology designs and compares favorably with other sequential response surface methodologies in the literature in terms of both design optimality and experimentation efficiency. Copyright

Mining Major Transitions of Chronic Conditions in Patients with Multiple Chronic Conditions

OBJECTIVES Evolution of multiple chronic conditions (MCC) follows a complex stochastic process, influenced by several factors including the inter-relationship of existing conditions, and patient-level risk factors. Nearly 20% of citizens aged 18 years and older are burdened with two or more (multiple) chronic conditions (MCC). Treatment for people living with MCC currently accounts for an estimated 66% of the Nations healthcare costs. However, it is still not known precisely how MCC emerge and accumulate among individuals or in the general population. This study investigates major patterns of MCC transitions in a diverse population of patients and identifies the risk factors affecting the transition process. METHODS A Latent regression Markov clustering (LRMCL) algorithm is proposed to identify major transitions of four MCC that include hypertension (HTN), depression, Post- Traumatic Stress Disorder (PTSD), and back pain. A cohort of 601,805 individuals randomly selected from the population of Iraq and Afghanistan war Veterans (IAVs) who received VA care during three or more years between 2002-2015, is used for training the proposed LRMCL algorithm. RESULTS Two major clusters of MCC transition patterns with 78% and 22% probability of membership respectively were identified. The primary cluster demonstrated the possibility of improvement when the number of MCC is small and an increase in probability of MCC accumulation as the number of co- morbidities increased. The second cluster showed stability (no change) of MCC overtime as the major pattern. Age was the most significant risk factor associated with the most probable cluster for each IAV. CONCLUSIONS These findings suggest that our proposed LRMCL algorithm can be used to describe and understand MCC transitions, which may ultimately allow healthcare systems to support optimal clinical decision- making. This method will be used to describe a broader range of MCC transitions in this and non-VA populations, and will add treatment information to see if models including treatments and MCC emergence can be used to support clinical decision-making in patient care.

A Truncated Logistic Regression Model in Probability of Detection Evaluation

ABSTRACT In nondestructive evaluation (NDE) studies, the probability of detection curve (POD) is an important performance metric. The traditional POD estimation is to conduct NDE inspections for artificially fabricated specimens with known flaws. This approach is often challenges because not only do fabricated flaws not adequately represent the flaws found in the field, but the cost and time of fabricating artificial specimens can also be very high. In practice, field samples and components in service with naturally occurring defects are readily available and much less expensive to test. However, the disadvantage of this field approach is that the exact number and sizes of the flaws in a sample (especially for flaws with small sizes) are unknown. As a result, serious bias in estimating the POD can occur. In this article, a truncated logistic regression method is developed that can estimate POD accurately and consistently with field samples based on multiple inspections. A case study illustrates the successful application of the proposed approach in a leading manufacturing company. The simulation studies also show that the proposed methods estimate the POD on field samples with quality comparable to that of the traditional approach on artificial specimens.

Mining patterns of comorbidity evolution in patients with multiple chronic conditions using unsupervised multi-level temporal Bayesian network

Over the past few decades, the rise of multiple chronic conditions has become a major concern for clinicians. However, it is still not known precisely how multiple chronic conditions emerge among patients. We propose an unsupervised multi-level temporal Bayesian network to provide a compact representation of the relationship among emergence of multiple chronic conditions and patient level risk factors over time. To improve the efficiency of the learning process, we use an extension of maximum weight spanning tree algorithm and greedy search algorithm to study the structure of the proposed network in three stages, starting with learning the inter-relationship of comorbidities within each year, followed by learning the intra-relationship of comorbidity emergence between consecutive years, and finally learning the hierarchical relationship of comorbidities and patient level risk factors. We also use a longest path algorithm to identify the most likely sequence of comorbidities emerging from and/or leading to specific chronic conditions. Using a de-identified dataset of more than 250,000 patients receiving care from the U.S. Department of Veterans Affairs for a period of five years, we compare the performance of the proposed unsupervised Bayesian network in comparison with those of Bayesian networks developed based on supervised and semi-supervised learning approaches, as well as multivariate probit regression, multinomial logistic regression, and latent regression Markov mixture clustering focusing on traumatic brain injury (TBI), post-traumatic stress disorder (PTSD), depression (Depr), substance abuse (SuAb), and back pain (BaPa). Our findings show that the unsupervised approach has noticeably accurate predictive performance that is comparable to the best performing semi-supervised and the second-best performing supervised approaches. These findings also revealed that the unsupervised approach has improved performance over multivariate probit regression, multinomial logistic regression, and latent regression Markov mixture clustering.

Sequential Laplacian Regularized V-Optimal Design of Experiments for Response Surface Modeling of Expensive Tests: An Application in Wind Tunnel Testing

Abstract In an increasing number of cases involving estimation of a response surface, one is often confronted with situations where there are several factors to be evaluated, but experiments are prohibitively expensive. In such scenarios, learning algorithms can actively query the user or other resources to determine the most informative settings to be tested. In this article, we propose an active learning methodology based on the fundamental idea of adding a ridge and a Laplacian penalty to the V-optimal design to shrink the weight of less significant factors, while looking for the most informative settings to be tested. To leverage the intrinsic geometry of the factor settings in highly nonlinear spaces, we generalize the proposed methodology to local regression. We also propose a simple sequential design strategy for efficient determination of subsequent experiments based on the information from previous experiments. The proposed methodology is particularly suited for problems involving expensive experiments with a high standard deviation of the error. We apply the proposed methodology to a simulated wind tunnel testing and compare the result with an existing practice. We also evaluate the estimation accuracy of the proposed methodology using the paper helicopter case study. Finally, through extensive simulated experiments, we demonstrate the performance of the proposed methodology against classic response surface methods in the literature.

A spatiotemporal outlier detection method based on partial least squares discriminant analysis and area Delaunay triangulation for image-based process monitoring

ABSTRACT Over the past two decades, statistical process control has evolved from monitoring individual data points to linear profiles to image data. Image sensors are now being deployed in complex systems at increasing rates due to the rich information they can provide. As a result, image data play an important role in process monitoring in different application domains ranging from manufacturing to service systems. Many of the existing process monitoring methods fail to take full advantage of the image data due to the datas complex nature in both the spatial and temporal domains. This article proposes a spatiotemporal outlier detection method based on the partial least squares discriminant analysis and a control statistic based on the area Delaunay triangulation of the squared prediction errors to improve the performance of an image-based monitoring scheme. First, the discriminant analysis of the partial least squares is used to efficiently extract the most important features from the high-dimensional image data to identify the benchmark images of the products and obtain the pixel value errors. Next, the squared errors resulting from the previous step are connected using a Delaunay triangulation to form a surface, the area of which is used as the control statistic for the purpose of outlier detection. A real case study at a paper product manufacturing company is used to compare the performance of the proposed method in detecting different types of outliers with some of the existing methods and demonstrate the merit of the proposed method.