A novel framework for prognostic factors identification of malignant mesothelioma through association rule mining

Analysis of nailfold capillaroscopy images with artificial intelligence: Data from literature and performance of machine learning and deep learning from images acquired in the SCLEROCAP study

OBJECTIVE

To evaluate the performance of machine learning and then deep learning to detect a systemic scleroderma (SSc) landscape from the same set of nailfold capillaroscopy (NC) images from the French prospective multicenter observational study SCLEROCAP.

METHODS

NC images from the first 100 SCLEROCAP patients were analyzed to assess the performance of machine learning and then deep learning in identifying the SSc landscape, the NC images having previously been independently and consensually labeled by expert clinicians. Images were divided into a training set (70 %) and a validation set (30 %). After features extraction from the NC images, we tested six classifiers (random forests (RF), support vector machine (SVM), logistic regression (LR), light gradient boosting (LGB), extreme gradient boosting (XGB), K-nearest neighbors (KNN)) on the training set with five different combinations of the images. The performance of each classifier was evaluated by the F1 score. In the deep learning section, we tested three pre-trained models from the TIMM library (ResNet-18, DenseNet-121 and VGG-16) on raw NC images after applying image augmentation methods.

RESULTS

With machine learning, performance ranged from 0.60 to 0.73 for each variable, with Hu and Haralick moments being the most discriminating. Performance was highest with the RF, LGB and XGB models (F1 scores: 0.75-0.79). The highest score was obtained by combining all variables and using the LGB model (F1 score: 0.79 ± 0.05, p < 0.01). With deep learning, performance reached a minimum accuracy of 0.87. The best results were obtained with the DenseNet-121 model (accuracy 0.94 ± 0.02, F1 score 0.94 ± 0.02, AUC 0.95 ± 0.03) as compared to ResNet-18 (accuracy 0.87 ± 0.04, F1 score 0.85 ± 0.03, AUC 0.87 ± 0.04) and VGG-16 (accuracy 0.90 ± 0.03, F1 score 0.91 ± 0.02, AUC 0.91 ± 0.04).

CONCLUSION

By using machine learning and then deep learning on the same set of labeled NC images from the SCLEROCAP study, the highest performances to detect SSc landscape were obtained with deep learning and in particular DenseNet-121. This pre-trained model could therefore be used to automatically interpret NC images in case of suspected SSc. This result nevertheless needs to be confirmed on a larger number of NC images.

Deep Learning-Based Microscopic Cell Detection Using Inverse Distance Transform and Auxiliary Counting

Microscopic cell detection is a challenging task due to significant inter-cell occlusions in dense clusters and diverse cell morphologies. This paper introduces a novel framework designed to enhance automated cell detection. The proposed approach integrates a deep learning model that produces an inverse distance transform-based detection map from the given image, accompanied by a secondary network designed to regress a cell density map from the same input. The inverse distance transform-based map effectively highlights each cell instance in the densely populated areas, while the density map accurately estimates the total cell count in the image. Then, a custom counting-aided cell center extraction strategy leverages the cell count obtained by integrating over the density map to refine the detection process, significantly reducing false responses and thereby boosting overall accuracy. The proposed framework demonstrated superior performance with F-scores of 96.93%, 91.21%, and 92.00% on the VGG, MBM, and ADI datasets, respectively, surpassing existing state-of-the-art methods. It also achieved the lowest distance error, further validating the effectiveness of the proposed approach. These results demonstrate significant potential for automated cell analysis in biomedical applications.

Diagnosis of Pleural Mesothelioma: Is Everything Solved at the Present Time?

Ranked high in worldwide growing health issues, pleural diseases affect approximately one million people globally per year and are often correlated with a poor prognosis. Among these pleural diseases, malignant pleural mesothelioma (PM), a neoplastic disease mainly due to asbestos exposure, still remains a diagnostic challenge. Timely diagnosis is imperative to define the most suitable therapeutic approach for the patient, but the choice of diagnostic modalities depends on operator experience and local facilities while bearing in mind the yield of each diagnostic procedure. Since the analysis of pleural fluid cytology is not sufficient in differentiating historical features in PM, histopathological and morphological features obtained via tissue biopsies are fundamental. The quality of biopsy samples is crucial and often requires highly qualified expertise. Since adequate tissue biopsy is essential, medical or video-assisted thoracoscopy (MT or VATS) is proposed as the most suitable approach, with the former being a physician-led procedure. Indeed, MT is the diagnostic gold standard for malignant pleural pathologies. Moreover, this medical or surgical approach can allow diagnostic and therapeutic procedures: it provides the possibility of video-assisted biopsies, the drainage of high volumes of pleural fluid and the administration of sterile calibrated talcum powder under visual control in order to achieve pleurodesis, placement of indwelling pleural catheters if required and in a near future potential intrapleural therapy. In this context, dedicated diagnostic pathways remain a crucial need, especially to quickly and properly diagnose PM. Lastly, the interdisciplinary approach and multidisciplinary collaboration should always be implemented in order to direct the patient to the best customised diagnostic and therapeutic pathway. At the present time, the diagnosis of PM remains an unsolved problem despite MDT (multidisciplinary team) meetings, mainly because of the lack of standardised diagnostic work-up. This review aims to provide an overview of diagnostic procedures in order to propose a clear strategy.

A data mining technique for detecting malignant mesothelioma cancer using multiple regression analysis

Lung cancer is a substantial health issue globally, and it is one of the main causes of mortality. Malignant mesothelioma (MM) is a common kind of lung cancer. The majority of patients with MM have no symptoms. In the diagnosis of any disease, etiology is crucial. MM risk factor detection procedures include positron emission tomography, magnetic resonance imaging, biopsies, X-rays, and blood tests, which are all necessary but costly and intrusive. Researchers primarily concentrated on the investigation of MM risk variables in the study. Mesothelioma symptoms were detected with the help of data from mesothelioma patients. The dataset, however, included both healthy and mesothelioma patients. Classification algorithms for MM illness diagnosis were carried out using computationally efficient data mining techniques. The support vector machine outperformed the multilayer perceptron ensembles (MLPE) neural network (NN) technique, yielding promising findings. With 99.87% classification accuracy achieved using 10-fold cross-validation over 5 runs, SVM is the best classification when contrasted to the MLPE NN, which achieves 99.56% classification accuracy. In addition, SPSS analysis is carried out for this study to collect pertinent and experimental data.

Data mining-based recommendation system using social networks-an analytical study

In the current age, social media is commonly used and shares enormous data. However, a huge amount of data makes it difficult to deal with. It requires a lot of storage and processing time. The content produced by social media needs to be stored efficiently by using data mining methods for providing suitable recommendations. The goal of the study is to perform a systematic literature review (SLR) which finds, analyzes, and evaluates studies that relate to data mining-based recommendation systems using social networks (DRSN) from 2011 to 2021 and open up a path for scientific investigations to enhance the development of recommendation systems in a social network. The SLR follows Kitchenhem's methodology for planning, guiding, and reporting the review. A systematic study selection procedure results in 42 studies that are analyzed in this article. The selected articles are examined on the base of four research questions. The research questions focus on publication venues, and chronological, and geographical distribution in DRSN. It also deals with approaches used to formulate DRSN, along with the dataset, size of the dataset, and evaluation metrics that validate the result of the selected study. Lastly, the limitations of the 42 studies are discussed. As a result, most articles published in 2018 acquired 21% of 42 articles, Whereas, China contributes 40% in this domain by comparing to other countries. Furthermore, 61% of articles are published in IEEE. Moreover, approximately 21% (nine out of 42 studies) use collaborative filtering for providing recommendations. Furthermore, the Twitter data set is common in that 19% of all other data sets are used, and precision and recall both cover 28% of selected articles for providing recommendations in social networks. The limitations show a need for a hybrid model that concatenates different algorithms and methods for providing recommendations. The study concludes that hybrid models may help to provide suitable recommendations on social media using data mining rules.

A Proposed Framework for Early Prediction of Schistosomiasis

Schistosomiasis is a neglected tropical disease that continues to be a leading cause of illness and mortality around the globe. The causing parasites are affixed to the skin through defiled water and enter the human body. Failure to diagnose Schistosomiasis can result in various medical complications, such as ascites, portal hypertension, esophageal varices, splenomegaly, and growth retardation. Early prediction and identification of risk factors may aid in treating disease before it becomes incurable. We aimed to create a framework by incorporating the most significant features to predict Schistosomiasis using machine learning techniques. A dataset of advanced Schistosomiasis has been employed containing recovery and death cases. A total data of 4316 individuals containing recovery and death cases were included in this research. The dataset contains demographics, socioeconomic, and clinical factors with lab reports. Data preprocessing techniques (missing values imputation, outlier removal, data normalisation, and data transformation) have also been employed for better results. Feature selection techniques, including correlation-based feature selection, Information gain, gain ratio, ReliefF, and OneR, have been utilised to minimise a large number of features. Data resampling algorithms, including Random undersampling, Random oversampling, Cluster Centroid, Near miss, and SMOTE, are applied to address the data imbalance problem. We applied four machine learning algorithms to construct the model: Gradient Boosting, Light Gradient Boosting, Extreme Gradient Boosting and CatBoost. The performance of the proposed framework has been evaluated based on Accuracy, Precision, Recall and F1-Score. The results of our proposed framework stated that the CatBoost model showed the best performance with the highest accuracy of (87.1%) compared with Gradient Boosting (86%), Light Gradient Boosting (86.7%) and Extreme Gradient Boosting (86.9%). Our proposed framework will assist doctors and healthcare professionals in the early diagnosis of Schistosomiasis.

Survival analysis and development of a prognostic nomogram for patients with malignant mesothelioma in different anatomic sites

Background

Malignant mesothelioma (MMe) is a rare and fatal cancer with a poor prognosis. Our study aimed to compare the overall survival (OS) of MMe patients across various sites and develop a prognostic model to provide a foundation for individualized management of MMe patients.

Methods

From the Surveillance, Epidemiology, and End Results (SEER) database, 1,772 individuals with malignant mesothelioma (MMe) were identified. The X-tile software was used to identify the optimal cut-off point for continuous variables. The Kaplan–Meier method was employed to compare the survival of MMe across different sites. The Cox proportional hazards model was applied to identify the independent risk factors of overall survival (OS) and a nomogram was constructed.

Results

In the survival analysis, MMe originating from the reproductive organs and hollow organs showed a relatively better prognosis than those originating from soft tissue, solid organs, and pleura. Age, gender, location, histological type, grade of differentiation, extent of disease, lymph node status, lymph node ratio (LNR), and chemotherapy were all found to be independent risk variables for the prognosis of MMe patients (P&lt;0.05) in a multivariate Cox analysis and were included in the construction of nomogram. In the training and testing sets, the C-index of the nomogram was 0.701 and 0.665, respectively, and the area under the ROC curve (AUROC) of the 1-, 3-, and 5-year overall survival rate was 0.749, 0.797, 0.833 and 0.730, 0.800, 0.832, respectively. The calibration curve shows that the nomogram is well-calibrated.

Conclusions

This is the first research to examine the prognosis of MMe patients based on the location. However, previous studies often focused on malignant pleural mesothelioma or malignant peritoneal mesothelioma with high incidence. Furthermore, a nomograph with good prediction efficiency was established according to the variables that influence patient survival outcomes, which provides us with a reference for clinical decision-making.

AIMIC: Deep Learning for Microscopic Image Classification

BACKGROUND AND OBJECTIVE

Deep learning techniques are powerful tools for image analysis. However, the lack of programming experience makes it difficult for novice users to apply this technology. This project aims to lower the barrier for clinical users to implement deep learning methods in microscopic image classification.

METHODS

In this study, an out-of-the-box software, AIMIC (artificial intelligence-based microscopy image classifier), was developed for users to apply deep learning technology in a code-free manner. The platform was equipped with state-of-the-art deep learning techniques and data preprocessing approaches. Furthermore, we evaluated the built-in networks on four benchmark microscopy image datasets to assist entry-level practitioners in selecting a suitable algorithm.

RESULTS

The entire deep learning pipeline, from training a new network to inferring unseen samples using the trained model, could be implemented on the proposed platform without the need for programming. In the evaluation experiments, the ResNeXt-50-32×4d outperformed other competitor algorithms in terms of average accuracy (96.83%) and average F1-score (96.82%). In addition, the MobileNet-V2 achieved a good balance between the performance (accuracy of 95.72%) and computational cost (inference time of 0.109s for identifying one sample).

CONCLUSIONS

The proposed AI platform allows people without programming experience to use artificial intelligence methods in microscopy image analysis. Besides, the ResNeXt-50-32×4d is a preferable solution for microscopic image classification, and MobileNet-V2 is most likely to be an alternative selection for the scenario when computing resources are limited.

Understanding the User-Generated Geographic Information by Utilizing Big Data Analytics for Health Care

There are two main ways to achieve an active lifestyle, the first is to make an effort to exercise and second is to have the activity as part of your daily routine. The study's major purpose is to examine the influence of various kinds of physical engagements on density dispersion of participants in Shanghai, China, and even prototype check-in data from a Location-Based Social Network (LBSN) utilizing a mix of spatial, temporal, and visualization methodologies. This paper evaluates Weibo used for big data evaluation and its dependability in some types rather than physically collected proofs by investigating the relationship between time, class, place, frequency, and place of check-in built on geographic features and related consequences. Kernel density estimation has been used for geographical assessment. Physical activities and frequency allocation are formed as a result of hour-to-day consumption habits. Our observations are based on customer check-in activities in physical venues such as gyms, parks, and playing fields, the prevalence of check-ins, peak times for visiting fun parks, and gender disparities, and we applied relative difference formulation to reveal the gender difference in a much better way. The purpose of this research is to investigate the influence of physical activity and health-related standard of living on well-being in a selection of Shanghai inhabitants.

Artificial Intelligence Techniques to Predict the Airway Disorders Illness: A Systematic Review

Airway disease is a major healthcare issue that causes at least 3 million fatalities every year. It is also considered one of the foremost causes of death all around the globe by 2030. Numerous studies have been undertaken to demonstrate the latest advances in artificial intelligence algorithms to assist in identifying and classifying these diseases. This comprehensive review aims to summarise the state-of-the-art machine and deep learning-based systems for detecting airway disorders, envisage the trends of the recent work in this domain, and analyze the difficulties and potential future paths. This systematic literature review includes the study of one hundred fifty-five articles on airway diseases such as cystic fibrosis, emphysema, lung cancer, Mesothelioma, covid-19, pneumoconiosis, asthma, pulmonary edema, tuberculosis, pulmonary embolism as well as highlights the automated learning techniques to predict them. The study concludes with a discussion and challenges about expanding the efficiency and machine and deep learning-assisted airway disease detection applications.

Service quality in football tourism: an evaluation model based on online reviews and data envelopment analysis with linguistic distribution assessments

The emergence of sports tourism has compelled sports managers to rethink the management and improvement of sports facilities. Through service quality analysis, sports managers can identify the strengths and weaknesses of their activities for possible advancement. Hence, this study aims to develop a decision support model based on integrating online reviews and data envelopment analysis to measure the degree of tourist satisfaction and provide benchmarking goals for service improvement. The proposed model employs text mining techniques to discover service quality attributes from text reviews. According to the discovered service quality attributes, we conduct sentiment analysis to reveal the sentiment polarities of the text reviews. Then, we refine the polarities and ratings of online reviews into linguistic distribution assessments. Furthermore, we develop a linguistic distribution output-oriented non-discretionary bestpoint slack-based measure (BP-SBM) to compute the degree of tourist satisfaction and benchmarking goals. The linguistic distribution output-oriented non-discretionary BP-SBM can handle both positive and negative data values, thus overcoming the flaws of the traditional model. Meanwhile, the proposed decision support model investigates how the service-quality attributes interact to provide improvement pathways for an underperforming stadium based on association rule mining. We test the applicability of the proposed decision support model on some Elite stadia in Europe.

Deep Ensemble Learning for the Automatic Detection of Pneumoconiosis in Coal Worker’s Chest X-ray Radiography

Globally, coal remains one of the natural resources that provide power to the world. Thousands of people are involved in coal collection, processing, and transportation. Particulate coal dust is produced during these processes, which can crush the lung structure of workers and cause pneumoconiosis. There is no automated system for detecting and monitoring diseases in coal miners, except for specialist radiologists. This paper proposes ensemble learning techniques for detecting pneumoconiosis disease in chest X-ray radiographs (CXRs) using multiple deep learning models. Three ensemble learning techniques (simple averaging, multi-weighted averaging, and majority voting (MVOT)) were proposed to investigate performances using randomised cross-folds and leave-one-out cross-validations datasets. Five statistical measurements were used to compare the outcomes of the three investigations on the proposed integrated approach with state-of-the-art approaches from the literature for the same dataset. In the second investigation, the statistical combination was marginally enhanced in the ensemble of multi-weighted averaging on a robust model, CheXNet. However, in the third investigation, the same model elevated accuracies from 87.80 to 90.2%. The investigated results helped us identify a robust deep learning model and ensemble framework that outperformed others, achieving an accuracy of 91.50% in the automated detection of pneumoconiosis.

Evolutionary Algorithm with Deep Auto Encoder Network Based Website Phishing Detection and Classification

Website phishing is a cyberattack that targets online users for stealing their sensitive data containing login credential and banking details. The phishing websites appear very similar to their equivalent legitimate websites for attracting a huge amount of Internet users. The attacker fools the user by offering the masked webpage as legitimate or reliable for retrieving its important information. Presently, anti-phishing approaches necessitate experts to extract phishing site features and utilize third-party services for phishing website detection. These techniques have some drawbacks, as the requirement of experts for extracting phishing features is time consuming. Many solutions for phishing websites attack have been presented, such as blacklist or whitelist, heuristics, and machine learning (ML) based approaches, which face difficulty in accomplishing effectual recognition performance due to the continual improvements of phishing technologies. Therefore, this study presents an optimal deep autoencoder network based website phishing detection and classification (ODAE-WPDC) model. The proposed ODAE-WPDC model applies input data pre-processing at the initial stage to get rid of missing values in the dataset. Then, feature extraction and artificial algae algorithm (AAA) based feature selection (FS) are utilized. The DAE model with the received features carried out the classification process, and the parameter tuning of the DAE technique was performed using the invasive weed optimization (IWO) algorithm to accomplish enhanced performance. The performance validation of the ODAE-WPDC technique was tested using the Phishing URL dataset from the Kaggle repository. The experimental findings confirm the better performance of the ODAE-WPDC model with maximum accuracy of 99.28%.

Design and Acceleration of Field Programmable Gate Array-Based Deep Learning for Empty-Dish Recycling Robots

As the proportion of the working population decreases worldwide, robots with artificial intelligence have been a good choice to help humans. At the same time, field programmable gate array (FPGA) is generally used on edge devices including robots, and it greatly accelerates the inference process of deep learning tasks, including object detection tasks. In this paper, we build a unique object detection dataset of 16 common kinds of dishes and use this dataset for training a YOLOv3 object detection model. Then, we propose a formalized process of deploying a YOLOv3 model on the FPGA platform, which consists of training and pruning the model on a software platform, and deploying the pruned model on a hardware platform (such as FPGA) through Vitis AI. According to the experimental results, we successfully realize acceleration of the dish detection using a YOLOv3 model based on FPGA. By applying different sparse training and pruning methods, we test the pruned model in 18 different situations on the ZCU102 evaluation board. In order to improve detection speed as much as possible while ensuring detection accuracy, for the pruned model with the highest comprehensive performance, compared to the original model, the comparison results are as follows: the model size is reduced from 62 MB to 12 MB, which is only 19% of the origin; the number of parameters is reduced from 61,657,117 to 9,900,539, which is only 16% of the origin; the running time is reduced from 14.411 s to 6.828 s, which is only less than half of the origin, while the detection accuracy is decreased from 97% to 94.1%, which is only less than 3%.

An In-Depth Survey of Bypassing Buffer Overflow Mitigation Techniques

Buffer Overflow (BOF) has been a ubiquitous security vulnerability for more than three decades, potentially compromising any software application or system. This vulnerability occurs primarily when someone attempts to write more bytes of data (shellcode) than a buffer can handle. To date, this primitive attack has been used to attack many different software systems, resulting in numerous buffer overflows. The most common type of buffer overflow is the stack overflow vulnerability, through which an adversary can gain admin privileges remotely, which can then be used to execute shellcode. Numerous mitigation techniques have been developed and deployed to reduce the likelihood of BOF attacks, but attackers still manage to bypass these techniques. A variety of mitigation techniques have been proposed and implemented on the hardware, operating system, and compiler levels. These techniques include No-EXecute (NX) and Address Space Layout Randomization (ASLR). The NX bit prevents the execution of malicious code by making various portions of the address space of a process inoperable. The ASLR algorithm randomly assigns addresses to various parts of the logical address space of a process as it is loaded in memory for execution. Position Independent Executable (PIE) and ASLR provide more robust protection by randomly generating binary segments. Read-only relocation (RELRO) protects the Global Offset Table (GOT) from overwriting attacks. StackGuard protects the stack by placing the canary before the return address in order to prevent stack smashing attacks. Despite all the mitigation techniques in place, hackers continue to be successful in bypassing them, making buffer overflow a persistent vulnerability. The current work aims to describe the stack-based buffer overflow vulnerability and review in detail the mitigation techniques reported in the literature as well as how hackers attempt to bypass them.

Rule-ranking method based on item utility in adaptive rule model

BACKGROUND

Decision-making is an important part of most human activities regardless of their daily activities, profession, or political inclination. Some decisions are relatively simple specifically when the consequences are insignificant while others can be very complex and have significant effects. Real-life decision problems generally involve several conflicting points of view (criteria) needed to be considered and this is the reason recent decision-making processes are usually supported by data as indicated by different data mining techniques. Data mining is the process of extracting data to obtain useful information and a promising and widely applied method is association rule mining which has the ability to identify interesting relationships between sets of items in a dataset and predict the associative behavior for new data. However, the number of rules generated in association rules can be very large, thereby making the exploitation process difficult. This means it is necessary to prioritize the selection of more valuable and relevant rules.

METHODS

Therefore, this study proposes a method to rank rules based on the lift ratio value calculated from the frequency and utility of the item. The three main functions in proposed method are mining of association rules from different databases (in terms of sources, characteristics, and attributes), automatic threshold value determination process, and prioritization of the rules produced.

RESULTS

Experiments conducted on six datasets showed that the number of rules generated by the adaptive rule model is higher and sorted from the largest lift ratio value compared to the apriori algorithm.

Knowledge Trajectories on Public Crisis Management Research from Massive Literature Text Using Topic-Clustered Evolution Extraction

Current research has ignored the hiddenness and the stochasticity of the evolution of public crisis management research, making the knowledge trajectories still unclear. This paper introduces a combined approach, LDA-HMM, to mine the hidden topics, present the evolutionary trajectories of the topics, and predict the future trends in the coming years to fill the research gaps. We reviewed 8543 articles in WOS from 1997 to 2021, extracted 39 hidden topics from the text using the LDA; 33 remained by manual labeling. The development of the topics over the years verifies that the topics are co-evolving with the public crisis events. The confusion and transition features indicate that most topics are confused or transferred to the others. The transition network and the direction of the topics show that six main transfer paths exist, and in the evolution process, the topics have become more focused. By training the HMM, we predict the trends in the next five years; the results show that the heat of the topic that focuses on traditional crisis issues will decrease while the focus on non-traditional issues will increase. We take the average error to test this model’s prediction effect by comparing it with the other approaches, concluding that it is better than the others. This study has practical implications for preventing crisis events, optimizing related policies, and grasping key research areas in the future.

Sustainable Network by Enhancing Attribute-Based Selection Mechanism Using Lagrange Interpolation

The security framework in Ad-hoc Networks (ANET) continues to attract the attention of researchers, although significant work has been accomplished already. Researchers in the last couple of years have shown quite an improvement in Identity Dependent Cryptography (IDC). Security in ANET is hard to attain due to the vulnerability of links (Wireless). IDC encompasses Polynomial Interpolations (PI) such as Lagrange, curve-fitting, and spline to provide security by implementing Integrated Key Management (IKM). The PI structure trusts all the available nodes in the network and randomly picks nodes for the security key generation. This paper presents a solution to the trust issues raised in Lagrange’s-PI (LI) utilizing an artificial neural network and attribute-based tree structure. The proposed structure not only improves the trust factor but also enhances the accuracy measures of LI to provide a sustainable network system. Throughput, PDR, noise, and latency have been increased by 47%, 50%, 34%, and 30%, respectively, by using LI and incorporating the aforementioned techniques.

An RG-FLAT-CRF Model for Named Entity Recognition of Chinese Electronic Clinical Records

The goal of Clinical Named Entity Recognition (CNER) is to identify clinical terms from medical records, which is of great importance for subsequent clinical research. Most of the current Chinese CNER models use a single set of features that do not consider the linguistic characteristics of the Chinese language, e.g., they do not use both word and character features, and they lack morphological information and specialized lexical information on Chinese characters in the medical field. We propose a RoBerta Glyce-Flat Lattice Transformer-CRF (RG-FLAT-CRF) model to address this problem. The model uses a convolutional neural network to discern the morphological information hidden in Chinese characters, and a pre-trained model to obtain vectors with medical features. The different vectors are stitched together to form a multi-feature vector. To use lexical information and avoid the problem of word separation errors, the model uses a lattice structure to add lexical information associated with each word, which can be used to avoid the problem of word separation errors. The RG-FLAT-CRF model scored 95.61%, 85.17%, and 91.2% for F1 on the CCKS 2017, 2019, and 2020 datasets, respectively. We used statistical tests to compare with other models. The results show that most p-values less than 0.05 are statistically significant.

Mining Campus Big Data: Prediction of Career Choice Using Interpretable Machine Learning Method

The issue of students’ career choice is the common concern of students themselves, parents, and educators. However, students’ behavioral data have not been thoroughly studied for understanding their career choice. In this study, we used eXtreme Gradient Boosting (XGBoost), a machine learning (ML) technique, to predict the career choice of college students using a real-world dataset collected in a specific college. Specifically, the data include information on the education and career choice of 18,000 graduates during their college years. In addition, SHAP (Shapley Additive exPlanation) was employed to interpret the results and analyze the importance of individual features. The results show that XGBoost can predict students’ career choice robustly with a precision, recall rate, and an F1 value of 89.1%, 85.4%, and 0.872, respectively. Furthermore, the interaction of features among four different choices of students (i.e., choose to study in China, choose to work, difficulty in finding a job, and choose to study aboard) were also explored. Several educational features, especially differences in grade point average (GPA) during their college studying, are found to have relatively larger impact on the final choice of career. These results can be of help in the planning, design, and implementation of higher educational institutions’ (HEIs) events.

Informative Power Evaluation of Clinical Parameters to Predict Initial Therapeutic Response in Patients with Advanced Pleural Mesothelioma: A Machine Learning Approach

Malignant pleural mesothelioma (MPM) is a rare neoplasm whose early diagnosis is challenging and systemic treatments are generally administered as first line in the advanced disease stage. The initial clinical response may represent a useful parameter in terms of identifying patients with a better long-term outcome. In this report, the initial therapeutical response in 46 patients affected with advanced/unresectable pleural mesothelioma was investigated. The initial therapeutic response was assessed by CT scan and clinical examination after 2–3 treatment cycles. Our preliminary evaluation shows that the group of patients treated with regimens including antiangiogenetics and/or immunotherapy had a significantly better initial response as compared to patients only treated with standard chemotherapy, exhibiting a disease control rate (DCR) of 100% (95% IC, 79.40–100%) and 80.0% (95% IC, 61.40–92.30%), respectively. Furthermore, the therapeutic response was correlated with the disease stage, blood leukocytes and neutrophils, high albumin serum levels, and basal body mass index (BMI). Specifically, the patients with disease stage III showed a DCR of 95.7% (95% IC, 78.1–99.9%), whereas for disease stage IV the DCR decreased to 66.7% (95% IC, 34.9–9.1%). Moreover, a better initial response was observed in patients with a higher BMI, who reached a DCR of 96.10% (95% IC, 80.36–99.90%). Furthermore, in order to evaluate in the predictive power of the collected features a multivariate way, we report the preliminary results of a machine learning model for predicting the initial therapeutic response. We trained a state-of-the-art algorithm combined to a sequential forward feature selection procedure. The model reached a median AUC value, accuracy, sensitivity, and specificity of 77.0%, 75%, 74.8%, and 83.3%, respectively. The features with greater informational power were gender, histotype, BMI, smoking habits, packs/year, and disease stage. Our preliminary data support the possible favorable correlation between innovative treatments and therapeutic response in patients with unresectable/advanced pleural mesothelioma. The small sample size does not allow concrete conclusions to be drawn; nevertheless, this work is the basis of an ongoing study that will also involve radiomics in a larger dataset.

Detecting Small Anatomical Structures in 3D Knee MRI Segmentation by Fully Convolutional Networks

Accurately identifying the pixels of small organs or lesions from magnetic resonance imaging (MRI) has a critical impact on clinical diagnosis. U-net is the most well-known and commonly used neural network for image segmentation. However, the small anatomical structures in medical images cannot be well recognised by U-net. This paper explores the performance of the U-net architectures in knee MRI segmentation to find a relative structure that can obtain high accuracies for both small and large anatomical structures. To maximise the utilities of U-net architecture, we apply three types of components, residual blocks, squeeze-and-excitation (SE) blocks, and dense blocks, to construct four variants of U-net, namely U-net variants. Among these variants, our experiments show that SE blocks can improve the segmentation accuracies of small labels. We adopt DeepLabv3plus architecture for 3D medical image segmentation by equipping SE blocks based on this discovery. The experimental results show that U-net with SE block achieves higher accuracy in parts of small anatomical structures. In contrast, DeepLabv3plus with SE block performs better on the average dice coefficient of small and large labels.

A Novel Method for Performance Measurement of Public Educational Institutions Using Machine Learning Models

Lack of education is a major concern in underdeveloped countries because it leads to poor human and economic development. The level of education in public institutions varies across all regions around the globe. Current disparities in access to education worldwide are mostly due to systemic regional differences and the distribution of resources. Previous research focused on evaluating students’ academic performance, but less has been done to measure the performance of educational institutions. Key performance indicators for the evaluation of institutional performance differ from student performance indicators. There is a dire need to evaluate educational institutions’ performance based on their disparities and academic results on a large scale. This study proposes a model to measure institutional performance based on key performance indicators through data mining techniques. Various feature selection methods were used to extract the key performance indicators. Several machine learning models, namely, J48 decision tree, support vector machines, random forest, rotation forest, and artificial neural networks were employed to build an efficient model. The results of the study were based on different factors, i.e., the number of schools in a specific region, teachers, school locations, enrolment, and availability of necessary facilities that contribute to school performance. It was also observed that urban regions performed well compared to rural regions due to the improved availability of educational facilities and resources. The results showed that artificial neural networks outperformed other models and achieved an accuracy of 82.9% when the relief-F based feature selection method was used. This study will help support efforts in governance for performance monitoring, policy formulation, target-setting, evaluation, and reform to address the issues and challenges in education worldwide.