Ling-Jing Kao

Background: The increasing rate of intensive care unit (ICU) readmissions poses significant challenges in healthcare, impacting both costs and patient outcomes. Predicting patient readmission after discharge is crucial for improving medical quality and reducing expenses. Traditional analyses of electronic health record (EHR) data have primarily focused on numerical data, often neglecting valuable text data. Methods: This study employs a hybrid model combining BERTopic and Long Short-Term Memory (LSTM) networks to predict ICU readmissions. Leveraging the MIMIC-III database, we utilize both quantitative and text data to enhance predictive capabilities. Our approach integrates the strengths of unsupervised topic modeling with supervised deep learning, extracting potential topics from patient records and transforming discharge summaries into topic vectors for more interpretable and personalized predictions. Results: Utilizing a comprehensive dataset of 36,232 ICU patient records, our model achieved an AUROC score of 0.80, thereby surpassing the performance of traditional machine learning models. The implementation of BERTopic facilitated effective utilization of unstructured data, generating themes that effectively guide the selection of relevant predictive factors for patient readmission prognosis. This significantly enhanced the model’s interpretative accuracy and predictive capability. Additionally, the integration of importance ranking methods into our machine learning framework allowed for an in-depth analysis of the significance of various variables. This approach provided crucial insights into how different input variables interact and impact predictions of patient readmission across various clinical contexts. Conclusions: The practical application of BERTopic technology in our hybrid model contributes to more efficient patient management and serves as a valuable tool for developing tailored treatment strategies and resource optimization. This study highlights the significance of integrating unstructured text data with traditional quantitative data to develop more accurate and interpretable predictive models in healthcare, emphasizing the importance of individualized care and cost-effective healthcare paradigms.

An ICU is a critical care unit that provides advanced medical support and continuous monitoring for patients with severe illnesses or injuries. Predicting the mortality rate of ICU patients can not only improve patient outcomes, but also optimize resource allocation. Many studies have attempted to create scoring systems and models that predict the mortality of ICU patients using large amounts of structured clinical data. However, unstructured clinical data recorded during patient admission, such as notes made by physicians, is often overlooked. This study used the MIMIC-III database to predict mortality in ICU patients. In the first part of the study, only eight structured variables were used, including the six basic vital signs, the GCS, and the patient’s age at admission. In the second part, unstructured predictor variables were extracted from the initial diagnosis made by physicians when the patients were admitted to the hospital and analyzed using Latent Dirichlet Allocation techniques. The structured and unstructured data were combined using machine learning methods to create a mortality risk prediction model for ICU patients. The results showed that combining structured and unstructured data improved the accuracy of the prediction of clinical outcomes in ICU patients over time. The model achieved an AUROC of 0.88, indicating accurate prediction of patient vital status. Additionally, the model was able to predict patient clinical outcomes over time, successfully identifying important variables. This study demonstrated that a small number of easily collectible structured variables, combined with unstructured data and analyzed using LDA topic modeling, can significantly improve the predictive performance of a mortality risk prediction model for ICU patients. These results suggest that initial clinical observations and diagnoses of ICU patients contain valuable information that can aid ICU medical and nursing staff in making important clinical decisions.

Cardiovascular diseases have been identified as one of the top three causes of death worldwide, with onset and deaths mostly due to heart failure (HF). In ICU, where patients with HF are at increased risk of death and consume significant medical resources, early and accurate prediction of the time of death for patients at high risk of death would enable them to receive appropriate and timely medical care. The data for this study were obtained from the MIMIC-III database, where we collected vital signs and tests for 6699 HF patient during the first 24 h of their first ICU admission. In order to predict the mortality of HF patients in ICUs more precisely, an integrated stacking model is proposed and applied in this paper. In the first stage of dataset classification, the datasets were subjected to first-level classifiers using RF, SVC, KNN, LGBM, Bagging, and Adaboost. Then, the fusion of these six classifier decisions was used to construct and optimize the stacked set of second-level classifiers. The results indicate that our model obtained an accuracy of 95.25% and AUROC of 82.55% in predicting the mortality rate of HF patients, which demonstrates the outstanding capability and efficiency of our method. In addition, the results of this study also revealed that platelets, glucose, and blood urea nitrogen were the clinical features that had the greatest impact on model prediction. The results of this analysis not only improve the understanding of patients’ conditions by healthcare professionals but allow for a more optimal use of healthcare resources.

In recent years, under the impact of digitization, all industries around the world have undergone unprecedented changes. Such changes have not only altered people’s consumption behavior but have also forced enterprises to accelerate the pace of digitization and actively start digital transformation. In this study, a literature review and focus group interview (FGI) were used to develop the dimensions and criteria to assess enterprise digital transformation status. To illustrate the digital transformation criteria proposed in this research, the retail industry was used as an example to measure the overall digital transformation performance by data envelopment analysis (DEA). The results show that the poor technical efficiency demonstrated by a vendor was not only due to the gradually decreasing returns to scale of the market; a decline in pure technical efficiency was also a contributor. In addition to adjusting their production on the basis of market conditions, vendors should properly manage their internal operations and pay attention to their resource and scale allocations to prevent reductions in their pure technical efficiency.

Predicting clinical patients’ vital signs is a leading critical issue in intensive care units (ICUs) related studies. Early prediction of the mortality of ICU patients can reduce the overall mortality and cost of complication treatment. Some studies have predicted mortality based on electronic health record (EHR) data by using machine learning models. However, the semi-structured data (i.e., patients’ diagnosis data and inspection reports) is rarely used in these models. This study utilized data from the Medical Information Mart for Intensive Care III. We used a Latent Dirichlet Allocation (LDA) model to classify text in the semi-structured data of some particular topics and established and compared the classification and regression trees (CART), logistic regression (LR), multivariate adaptive regression splines (MARS), random forest (RF), and gradient boosting (GB). A total of 46,520 ICU Patients were included, with 11.5% mortality in the Medical Information Mart for Intensive Care III group. Our results revealed that the semi-structured data (diagnosis data and inspection reports) of ICU patients contain useful information that can assist clinical doctors in making critical clinical decisions. In addition, in our comparison of five machine learning models (CART, LR, MARS, RF, and GB), the GB model showed the best performance with the highest area under the receiver operating characteristic curve (AUROC) (0.9280), specificity (93.16%), and sensitivity (83.25%). The RF, LR, and MARS models showed better performance (AUROC are 0.9096, 0.8987, and 0.8935, respectively) than the CART (0.8511). The GB model showed better performance than other machine learning models (CART, LR, MARS, and RF) in predicting the mortality of patients in the intensive care unit. The analysis results could be used to develop a clinically useful decision support system.

With the development of information technology, online transactions and e‐commerce are gradually replacing conventional consumption patterns. To obtain a competitive advantage, industries proactively engage in digital transformations and the management of e‐commerce platforms. Faced with changes in market patterns, e‐commerce channels and online advertising firms hope to extend users' website browsing duration/time on site to enhance the effects of product promotion and the likelihood of advertisement clicks. The greatest challenge in predicting time on site is that clickstream data are not mutually independent, and short‐, mid‐, and long‐term data may intervene in a time series. Such timing dependence increases difficulty of capturing or learning the characteristics of website users for ordinary prediction models and leads to confusion and deviation during model construction. Accordingly, this study proposed a prediction method integrating self‐organizing map (SOM) and long short‐term memory (LSTM). The SOM method was initially applied to categorize website members into groups based on similarities in browsing behavior, and the LSTM prediction model was subsequently developed using the webpage browsing data of each group. The performance of the proposed method is evaluated by comparing the prediction with the results of three competing approaches (SOM with support vector regression, SOM with multilayer perceptron, and single LSTM) on the clickstream data provided by a leading online retailer specializing in selling skin care and cosmetics products in Taiwan. The Wilcoxon signed‐rank test validated the proposed SOM‐LSTM model outperforms competing approaches in remaining time‐on‐site prediction. This study serves as a first attempt to systematically predict remaining time on site for e‐commerce users in terms of empirically verifying a hybrid approach which integrates SOM and LSTM techniques.

According to previous studies, the workforce competencies of engineers in manufacturing industries generally comprise technical, generic, and knowledge competencies. However, these categories of workforce competencies are insufficient for firms whose manufacturing processes are very easily interrupted by anomalies and whose core competency is highly related to production yield. In addition to drawing attention to the importance of identifying engineers’ problem-solving competencies, this article proposes a procedure to identify and rate workforce competencies for firms in the manufacturing industry. An IC packaging process engineer was used as an example to illustrate the proposed competency identification and rating procedure. In contrast to previous studies, this article suggests that problem-solving competencies can be identified from engineers’ responses to the anomalies commonly encountered in each step of the manufacturing process and that operational competency can be identified from the knowledge, techniques, or capabilities required to fulfill a task in each step of the manufacturing process. For the IC packaging and testing industry, this article can help human resources not only evaluate the competency of job candidates but also design training courses according to the job tenure of engineers. Furthermore, this article results can be used to design teaching material for apprenticeship to help engineers develop competencies in a structured manner and enhance their sensitivities to production-line problems.

Companies often extend product lines with the goal of increasing demand for their products and responding to competitive threats. Although line extensions may lead to cannibalization and reduction of overall profit, the bulk of theoretical and empirical research has suggested that product line extensions result in a net gain of overall demand and market share. To mitigate cannibalization, the extant literature prescribes the addition of premium versions of products, or “upward line extensions,” with the intention of achieving gains not only in demand and market share but also in overall profit. In this research, the authors employ analytical and empirical methods to make the case that upward line extensions aimed at matching a competing product's attribute may lead consumers to reassess their perceptions about the brand and the attributes of products in the market in a way that erodes the advantages of the extending firm. Ultimately, this can result in a loss of demand, market share, and profit for the extending firm.

The purpose of this study is to evaluate the training institution performance and to improve the management of the Manpower Training Project (MTP) administered by the Semiconductor Institute in Taiwan. Much literature assesses the efficiency of an internal training program initiated by a firm, but only little literature studies the efficiency of an external training program led by government. In the study, a hybrid solution of ICA-DEA and ICA-MPI is developed for measuring the efficiency and the productivity growth of each training institution over the period. The technical efficiency change, the technological change, pure technical efficiency change, scale efficiency change, and the total factor productivity change were evaluated according to five inputs and two outputs. According to the results of the study, the training institutions can be classified by their efficiency successfully and the guidelines for the optimal level of input resources can be obtained for each inefficient training institution. The Semiconductor Institute in Taiwan can allocate budget more appropriately and establish withdrawal mechanisms for inefficient training institutions.