Text classification is an essential aspect in many applications, such as spam detection and sentiment analysis. With the growing number of textual documents and datasets generated through social media and news articles, an increasing number of machine learning methods are required for accurate textual classification. For this paper, a comprehensive evaluation of the performance of multiple supervised learning models, such as logistic regression (LR), decision trees (DT), support vector machine (SVM), AdaBoost (AB), random forest (RF), multinomial naive Bayes (NB), multilayer perceptrons (MLP), and gradient boosting (GB), was conducted to assess the efficiency and robustness, as well as limitations, of these models on the classification of textual data. SVM, LR, and MLP had better performance in general, with SVM being the best, while DT and AB had much lower accuracies amongst all the tested models. Further exploration on the use of different SVM kernels was performed, demonstrating the advantage of using linear kernels over polynomial, sigmoid, and radial basis function kernels for text classification. The effects of removing stop words on model performance was also investigated; DT performed better with stop words removed, while all other models were relatively unaffected by the presence or absence of stop words.

Failure modes and effects analysis ( $FMEA$ ) substantially facilitates the efforts of industrial manufacturers in prioritizing failures that require corrective actions to continuously improve product quality. However, the conventional approach fails to provide satisfactory explanation of the aggregate effects of a failure from different perspectives such as technical severity, economic severity, and production capacity in some practical applications. To fulfill the existing gap in the $FMEA$ literature, this paper proposes an extension by considering associated quality cost and the capability of failure detection system as additional determinants to signify the priority level for each failure mode. The quality cost and capacity are considered as key factors for sustainable survival and development of an industrial manufacturer in the fierce competition market these days. The performance of the extended scheme was tested in an empirical case at a non-woven fabrics manufacturer. Analytical results indicate that the proposed approach outperforms the traditional one and remarkably reduces the percentage of defective fabrics from about 2.41% before the trial period to 1.13%,thus significantly reducing wastes and increasing operation efficiency, thereby providing valuable advantages to improve organizational competition power for their sustainable growth.

Assistive braille technology has existed for many years with the purpose of aiding the blind in performing common tasks such as reading, writing, and communicating with others. Such technologies are aimed towards helping those who are visually impaired to better adapt to the visual world. However, an obvious gap exists in current technology when it comes to symmetric two-way communication between the blind and non-blind, as little technology allows non-blind individuals to understand the braille system. This research presents a novel approach to convert images of braille into English text by employing a convolutional neural network (CNN) model and a ratio character segmentation algorithm (RCSA). Further, a new dataset was constructed, containing a total of 26,724 labeled braille images, which consists of 37 braille symbols that correspond to 71 different English characters, including the alphabet, punctuation, and numbers. The performance of the CNN model yielded a prediction accuracy of 98.73% on the test set. The functionality performance of this artificial intelligence (AI) based recognition system could be tested through accessible user interfaces in the future.

Reducing the warranty cost to increase profit is a major focus for most businesses. To achieve this goal, the acceptance sampling plan (ASP) can be an efficient tool to validate whether the product lifetimes meet their requirements. In the ASP practice, the single sampling plan (SSP) is widely used due to its administrative simplicity. Unfortunately, the SSP ignores the valuable sample information from preceding lots, which may reduce its efficiency and quality-discrimination power. Especially, the collection of lifetime information often consumes a significant amount of time and is expensive in today’s high product-yield and long-life environment. To address the drawbacks of the SSP, we propose a lifetime performance-qualified quick switch sampling (QSS) system under a Weibull distribution with failure-censoring. This system nimbly switches between the normal SSP and tightened SSP based on the previous lot-disposition result. Compared to the ordinary SSP and the recently proposed multiple-lot dependent sampling plan, the QSS system not only significantly reduces the required number of failures that need to be observed in the life testing but also exhibits a better discriminatory operating characteristic curve for the lot disposition. On the other hand, the QSS system helps the consumer to communicate to the supplier important messages that (i) only with reliable submissions the latter can persistently enjoy the benefits of the normal lot-sentencing standard; (ii) once the submitted lot is rejected, the supplier will have to spend more efforts to gain back the consumer trust. Moreover, we developed a web-based tool to help practitioners obtain the system criteria easily and quickly without the need for the traditional time-consuming table lookup. By operating the interactive interface of the web app, practitioners can quickly obtain the system criteria conforming to the requirements. Finally, the proposed QSS system is illustrated by a real-life application.

Acceptance sampling plans (ASPs), practically providing an economic size of sampling inspection and a definite standard for lot-sentencing determination, have been widely used for the products' quality verification and validation. Simply sentencing the submitted lot according to the inspection of only one group of samples, the single sampling plan (SSP) has an advantage of lot-by-lot simple administration. When encountering costly, destructive, and time-consuming life testing, the SSP’s simplicity, however, comes at the high expense of requiring large-tested samples. To overcome the drawback, the multiple-dependent-state sampling plan (MDSP) and repetitive group sampling plan (RGSP) have been recently developed. The former activates a backtracking mechanism to include valuable information inspected from preceding lots, whereas the latter actuates a resampling mechanism to alleviate psychological skepticism raised from both entities of supplier and buyer. In this paper, an integrated failure-censored sampling scheme (IFCSS), consolidating both advantages gained from the MDSP and the RGSP, is developed based on the Weibull-lifetime performance index. We construct an optimization model to determine the plan criteria under the economical objective. Moreover, we further create a Weibull-lifetime IFCSS web-based app with QR code access for the plan’s easy setup. By comparing with inspection performances, our proposed IFCSS exhibits superior cost-efficiency and discriminating capability than the recently developed MDSP and RGSP. Finally, the Weibull-lifetime IFCSS’s applicability is illustrated by a ceramic-bearing lifetime-testing inspection.

The quick-switch sampling (QSS) systems based on process capability index (PCI) have been proved to be useful for inspecting processes streaming at a low level of defects because of their small sizes required and dynamic responsiveness to problems. However, most of the recent studies have aimed at deriving the mathematical model and comparing performances for the acceptance criteria-type QSS system, and only a few of them have an interest in the development of the required sample-size-type QSS system; their studies have limited simultaneous discussions on both systems’ advantages and disadvantages, not to mention the lack of investigations on the impact of the joint application of the two QSS system types. In this paper, an integrated QSS (IQSS) system based on the most popular PCI was proposed; the system can accommodate two existing types of PCI-based QSS systems and further boost the sampling performance and discriminatory power. Moreover, by operating suitable types of PCI-based QSS systems in different stages of the supplier-buyer partnership, a progressive lot-disposition strategy was introduced to construct a solid supplier-buyer relationship. We also developed a web-based tool to accurately and efficiently execute all types’ PCI-based QSS systems. Finally, the industrial applicability was demonstrated in a case study.