scholarly journals Using Principal Component Analysis and Artificial Neural Networks for Fault Type Forecasting in an Automotive Company

2018 ◽  
Vol 1 (1) ◽  
pp. 1110-1119
Author(s):  
Tülay Korkusuz Polat
Keyword(s):  
Neural Networks ◽  
Principal Component Analysis ◽  
Quality Control ◽  
Principal Component ◽  
Production Quality ◽  
Quality Score ◽  
The Other ◽  
Paint Shop ◽  
Fault Type ◽  
Automotive Company

In this study, failures that occurred in the paint shop of an automotive company were discussed. The relationships between these failures and the probabilities of prospective occurrences were investigated. Any product produced in the company passes quality control at the end of production. Technical or operator-originated types of potential failures are examined during this control. Causes of failures in the paint shop and how they can be resolved pose a serious problem, just as in the other departments of the factory. This is because every failure encountered negatively influence the product quality and harm the company in terms of cost/productivity/image. The inability of the paint shop to predict the probability of failures in advance and its inability to establish a link between the types of failures also lead to its failure to pass quality control — which is the subsequent process — and cause its “production quality score” to fall, as well as other adversities. This study was carried out to determine which failures were usually caused by the activities in the paint shop and to develop a model that would predict the pass/fail state of the types of failures in question.

scholarly journals Convolutional Neural Networks with Transfer Learning for Recognition of COVID-19: A Comparative Study of Different Approaches

AI ◽  
2020 ◽  
Vol 1 (4) ◽  
pp. 586-606
Author(s):  
Tanmay Garg ◽  
Mamta Garg ◽  
Om Prakash Mahela ◽  
Akhil Ranjan Garg
Keyword(s):  
Neural Networks ◽  
Principal Component Analysis ◽  
Feature Selection ◽  
Convolutional Neural Networks ◽  
Image Representation ◽  
Principal Component ◽  
Classification Problem ◽  
The Other ◽  
X Ray ◽  
Image Representations

To judge the ability of convolutional neural networks (CNNs) to effectively and efficiently transfer image representations learned on the ImageNet dataset to the task of recognizing COVID-19 in this work, we propose and analyze four approaches. For this purpose, we use VGG16, ResNetV2, InceptionResNetV2, DenseNet121, and MobileNetV2 CNN models pre-trained on ImageNet dataset to extract features from X-ray images of COVID and Non-COVID patients. Simulations study performed by us reveal that these pre-trained models have a different level of ability to transfer image representation. We find that in the approaches that we have proposed, if we use either ResNetV2 or DenseNet121 to extract features, then the performance of these approaches to detect COVID-19 is better. One of the important findings of our study is that the use of principal component analysis for feature selection improves efficiency. The approach using the fusion of features outperforms all the other approaches, and with this approach, we could achieve an accuracy of 0.94 for a three-class classification problem. This work will not only be useful for COVID-19 detection but also for any domain with small datasets.

scholarly journals Principal component analysis on the perceptions of milk producers about Rhipicephalus (Boophilus) microplus control in Minas Gerais

2012 ◽  
Vol 21 (3) ◽  
pp. 224-231 ◽  
Author(s):  
Christiane Maria Barcellos Magalhães da Rocha ◽  
Fábio Raphael Pascoti Bruhn ◽  
Romário Cerqueira Leite ◽  
Antônio Marcos Guimarães ◽  
Ivan Barbosa Sampaio ◽  
...  
Keyword(s):  
Principal Component Analysis ◽  
Principal Component ◽  
Component Analysis ◽  
Production Quality ◽  
Minas Gerais ◽  
Boophilus Microplus ◽  
Basic Knowledge ◽  
Positive Correlation ◽  
Perceptions And Attitudes ◽  
Technological Level

Milk producers in Lavras, Passos and Divinópolis, Minas Gerais, were interviewed with the aim of evaluating their perceptions and attitudes regarding control over Rhipicephalus (Boophilus) microplus. Multivariate correlation between the variables was done by means of principal component analysis. The producers' perceptions and attitudes regarding R. (B.) microplus were similar: most of them did not have any basic knowledge of tick biology or control, and they applied acaricide products through backpack spraying, without any defined technical criteria. Some of the results obtained were: I. a negative correlation between schooling level and the frequency of spraying cattle with acaricides; II. a positive correlation between milk production, quality of equipment for acaricide application and technological level of the farm; III. farm properties with greater production and technification tended to keep the intervals between acaricide applications constant over the course of the year. After applying principal component analysis, a positive correlation was observed between schooling level, technological level of the farm and perceptions regarding R. (B.) microplus, but without any correlation with attitudes towards controlling this tick. It was concluded that higher technological level and schooling level improved the producers' perceptions relating to the biology of the tick R. (B.) microplus, but did not achieve effectiveness with regard to using controls more rationally.

Correlating Variations in the Dynamic Resistance Signature to Weld Strength in Resistance Spot Welding Using Principal Component Analysis

2016 ◽  
Vol 139 (4) ◽  
Author(s):  
David W. Adams ◽  
Cameron D. E. Summerville ◽  
Brendan M. Voss ◽  
Jack Jeswiet ◽  
Matthew C. Doolan
Keyword(s):  
Principal Component Analysis ◽  
Quality Control ◽  
Spot Welding ◽  
Correlation Coefficients ◽  
Principal Component ◽  
Weld Strength ◽  
Dynamic Resistance ◽  
Resistance Spot ◽  
Resistance Spot Welds ◽  
Key Features

Traditional quality control of resistance spot welds by analysis of the dynamic resistance signature (DRS) relies on manual feature selection to reduce the dimensionality prior to analysis. Manually selected features of the DRS may contain information that is not directly correlated to strength, reducing the accuracy of any classification performed. In this paper, correlations between the DRS and weld strength are automatically detected by calculating correlation coefficients between weld strength and principal components of the DRS. The key features of the DRS that correlate to weld strength are identified in a systematic manner. Systematically identifying relevant features of the DRS is useful as the correlations between weld strength and DRS may vary with process parameters.

Sign in / Sign up

Export Citation Format

Share Document