A unique solution for principal component analysis-based multi-response optimization problems

Fard, Nasser; Xu, Huyang; Fang, Yuanchen

doi:10.1007/s00170-015-7358-x

Cited by 9 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, as [96] explained, highly correlated features lead to overfitting, and the PCA technique is applied to remove the highly correlated features based on the correlation matrix, thus increasing the prediction accuracy. Without increasing the computational complexity, it considers potential correlations between the answer variables [97], transforming the data into uncorrelated features that assist in converting the data from a high-dimensional to a low-dimensional space, retaining the maximum amount of information [98]. As there were large differences between the ranges of the variables that were provided as inputs to the PCA, a normalisation of the data was first performed.…”

Section: Methodsmentioning

confidence: 99%

Predicting the Health Status of a Pulp Press Based on Deep Neural Networks and Hidden Markov Models

et al. 2023

View full text Add to dashboard Cite

The maintenance paradigm has evolved over the last few years and companies that want to remain competitive in the market need to provide condition-based maintenance (CBM). The diagnosis and prognosis of the health status of equipment, predictive maintenance (PdM), are fundamental strategies to perform informed maintenance, increasing the company’s profit. This article aims to present a diagnosis and prognosis methodology using a hidden Markov model (HMM) classifier to recognise the equipment status in real time and a deep neural network (DNN), specifically a gated recurrent unit (GRU), to determine this same status in a future of one week. The data collected by the sensors go through several phases, starting by cleaning them. After that, temporal windows are created in order to generate statistical features of the time domain to better understand the equipment’s behaviour. These features go through a normalisation to produce inputs for a feature extraction process, via a principal component analysis (PCA). After the dimensional reduction and obtaining new features with more information, a clustering is performed by the K-means algorithm, in order to group similar data. These clusters enter the HMM classifier as observable states. After training using the Baum–Welch algorithm, the Viterbi algorithm is used to find the best path of hidden states that represent the diagnosis of the equipment, containing three states: state 1—“State of Good Operation”; state 2—“Warning State”; state 3—“Failure State”. Once the equipment diagnosis is complete, the GRU model is used to predict the future, both of the observable states as well as the hidden states coming out from the HMM. Thus, through this network, it is possible to directly obtain the health states 7 days ahead, without the necessity to run the whole methodology from scratch.

show abstract

Section: Methodsmentioning

confidence: 99%

Predicting the Health Status of a Pulp Press Based on Deep Neural Networks and Hidden Markov Models

et al. 2023

View full text Add to dashboard Cite

show abstract

“…After obtaining each response's value, we apply the WPCA method to analyze. The WPCA method is a method that first preprocesses different categories of sample data, then calculates weights using the feature relationships between different categories in the data, weights the data samples, and later performs feature extraction using PCA [16,17,18]. In this article, we apply the WPCA approach to the data of the three response variables.…”

Section: Simulation Experimentsmentioning

confidence: 99%

Simulation-Based Multi-Response Optimization for Primary Healthcare System Design and Operation

Qin,

Yang,

et al. 2023

Advances in Economics, Business and Management Research

Self Cite

View full text Add to dashboard Cite

Primary care physician service is a good form of medical service. Many countries, such as China, are in the early stages of developing family doctor services and need a mature service system. The services residents receive after signing up with a primary care physician must be clarified. In many urban communities in China, universal healthcare policies have emerged. However, most need more resources, and primary care physicians and nurses cannot provide primary care services to all community residents. This study designs a primary health care system involving general practitioners and nurses based on a Chengdu, Sichuan Province community. Moreover, the operation of this primary healthcare system is modeled by a discrete event simulation to find the service capability of the system. Then we design the optimization experiments using Taguchi's method, define multiple factors and corresponding levels, simulate the system operation by simulation model for different combinations of factors and levels, and obtain the corresponding response variables. Finally, we apply the weighted principal component analysis based on multi-response optimization to find the optimal human resource allocation, which provides guidance for better service of the primary health care system.

show abstract

“…However, those are not easily comprehendible and thus, are quite difficult to implement in industry. Keeping this in mind, a group of researchers (Tong and Hsieh 2000;Chiang and Su 2003;Tong et al 2005;Pan et al 2007;Pal and Gauri, 2010;Khanna et al, 2015;Priyadarshini et al, 2015;Fard, et al, 2016;Ghani et al, 2017) have proposed some simplified approaches for the multi-response optimization, which can be applied under the framework of Taguchi method (1986) of experimentation and analysis. In these approaches, the quality loss or SNR of individual responses is converted first into an overall process performance index (PPI) and then, the optimal settings of the input variables are determined by examining the level averages on the PPI.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Among various approaches for the multi-response optimization, the desirability function-based approach (Derringer and Suich, 1980) has gained the maximum popularity in solving the multiresponse optimization problems.Subsequently, many researchers (Castillo et al, 1996;Kim and Lin, 2000;Jeong and Kim, 2003;Wu, 2005;Kushwaha et al, 2013;Candioti et al, 2014;Salmasnia and Bashiri, 2015;Ahmad et al, 2017) modified the desirability functions and used it for optimization of multiple quantitative response variables.All these approaches are mathematically rigorous and consequently, are not easy to implement in industries. Several other researchers, therefore, have taken interest in developing appropriate methodologies for optimizing multiple responses under Taguchi's robust design framework (Tong and Hsieh 2000;Chiang and Su 2003;Tong et al 2005;Pan et al 2007;Pal and Gauri, 2010;Khanna et al, 2015;Priyadarshini et al, 2015;Fard, et al, 2016;Ghani et al, 2017). However, all these researchers make an implicit assumption that all the responses are quantitative variables and thus, their proposed methodologies for multi-response optimization are applicable only if all the response variables are quantitative in nature.…”

Section: Introductionmentioning

confidence: 99%

A desirability functions-based approach for simultaneous optimization of quantitative and ordinal response variables in industrial processes

Pal¹,

Gauri²

2018

Int. J. Eng. Sci. Tech

View full text Add to dashboard Cite

The most important step for producing high quality products is the optimum utilization of the manufacturing processes and its resources, which can be accomplished by using optimal process settings. Extensive research works are reported in literature for optimization of process settings with respect to single as well as multiple quantitative response variables. However, in real world, often some aspects of product quality (e.g. dimensions, yield etc.) are measured quantitatively and some other aspects of product quality (e.g. color, finish etc.) are assessed qualitatively like 'poor', 'average', 'good' etc., which can be treated as ordered categorical or ordinal variables. Only a few researchers have attempted to tackle the problem of simultaneous optimization of quantitative and ordinal response variables. But none of these approaches results in an efficient optimal solution with respect to the desirability values of all the individual response variables. It motivated us to develop a more useful approach for simultaneously optimizing industrial processes with respect to quantitative as well as ordinal response variables. In the proposed method, the concept of desirability functions for continuous quantitative variables is extended for ordinal response variables, and then process settings is optimized considering geometric mean of the individual desirability functions of continuous as well as ordinal response variables. A set of experimental data, which have been analyzed by the past researchers, are analyzed using the proposed method and the related results are compared. The comparison of results reveals that the proposed method leads to better optimal solution than the other methods.

show abstract

A unique solution for principal component analysis-based multi-response optimization problems

Cited by 9 publications

References 9 publications

Predicting the Health Status of a Pulp Press Based on Deep Neural Networks and Hidden Markov Models

Predicting the Health Status of a Pulp Press Based on Deep Neural Networks and Hidden Markov Models

Simulation-Based Multi-Response Optimization for Primary Healthcare System Design and Operation

A desirability functions-based approach for simultaneous optimization of quantitative and ordinal response variables in industrial processes

Contact Info

Product

Resources

About