Original Dataset Research Articles

Decision-Making on the Diagnosis of Oncological Diseases Using Cost-Sensitive SVM Classifiers Based on Datasets with a Variety of Features of Different Natures

This paper discusses the problem of detecting cancer using such biomarkers as blood protein markers. The purpose of this research is to propose an approach for making decisions in the diagnosis of cancer through the creation of cost-sensitive SVM classifiers on the basis of datasets with a variety of features of different nature. Such datasets may include compositions of known features corresponding to blood protein markers and new features constructed using methods for calculating entropy and fractal dimensions, as well as using the UMAP algorithm. Based on these datasets, multiclass SVM classifiers were developed. They use cost-sensitive learning principles to overcome the class imbalance problem, which is typical for medical datasets. When implementing the UMAP algorithm, various variants of the loss function were considered. This was performed in order to select those that provide the formation of such new features that ultimately allow us to develop the best cost-sensitive SVM classifiers in terms of maximizing the mean value of the metric MacroF1−score. The experimental results proved the possibility of applying the UMAP algorithm, approximate entropy and, in addition, Higuchi and Katz fractal dimensions to construct new features using blood protein markers. It turned out that when working with the UMAP algorithm, the most promising is the application of a loss function on the basis of fuzzy cross-entropy, and the least promising is the application of a loss function on the basis of intuitionistic fuzzy cross-entropy. Augmentation of the original dataset with either features on the basis of the UMAP algorithm, features on the basis of the UMAP algorithm and approximate entropy, or features on the basis of approximate entropy provided the creation of the three best cost-sensitive SVM classifiers with mean values of the metric MacroF1−score increased by 5.359%, 5.245% and 4.675%, respectively, compared to the mean values of this metric in the case when only the original dataset was utilized for creating the base SVM classifier (without performing any manipulations to overcome the class imbalance problem, and also without introducing new features).

LifeWatch ERIC: papers collection on original datasets and new e-services for the biodiversity and ecosystems' scientific community.

Papers including articles that are produced because of the activities of LifeWatch ERIC, in the context of its second implementation period (2022 - 2026) and through the implementation of its new Strategic Working Plan, are published in this special collection. The articles include data papers, papers describing the development and functioning of analytical services and papers describing any other research outcome, produced either by LifeWatch ERIC or by any collaboration with any other ERIC, Research Infrastructure, global aggregator or other legal entity.

Lightweight YOLOv7 Algorithm for Multi-Object Recognition on Contrabands in Terahertz Images

With the strengthening of worldwide counter-terrorism initiatives, it is increasingly important to detect contrabands such as controlled knives and flammable materials hidden in clothes and bags. Terahertz (THz) imaging technology is widely used in the field of contraband detection due to its advantages of high imaging speed and strong penetration. However, the terahertz images are of poor qualities and lack texture details. Traditional target detection methods suffer from low detection speeds, misdetection, and omission of contraband. This work pre-processes the original dataset using a variety of image processing methods and validates the effect of these methods on the detection results of YOLOv7. Meanwhile, the lightweight and multi-object detection YOLOv7 (LWMD-YOLOv7) algorithm is proposed. Firstly, to meet the demand of real-time for multi-target detection, we propose the space-to-depth mobile (SPD_Mobile) network as the lightweight feature extraction network. Secondly, the selective attention module large selective kernel (LSK) network is integrated into the output of the multi-scale feature map of the LWMD-YOLOv7 network, which enhances the effect of feature fusion and strengthens the network’s attention to salient features. Finally, Distance Intersection over Union (DIOU) is used as the loss function to accelerate the convergence of the model and to have a better localisation effect for small targets. The experimental results show that the YOLOv7 algorithm achieves the best detection results on the terahertz image dataset after the non-local mean filtering process. The LWMD-YOLOv7 algorithm achieves a detection accuracy P of 98.5%, a recall R of 97.5%, and a detection speed of 112.4 FPS, which is 26.9 FPS higher than that of the YOLOv7 base network. The LWMD-YOLOv7 achieves a better balance between detection accuracy and detection speed. It provides a technological reference for the automated detection of contraband in terahertz images.

The competence‐control dilemma and the institutional design of European Union agencies

AbstractIn the past decades, European Union (EU) agencies have proliferated to address a plethora of governance problems. When designing EU agencies, EU legislators confront a tension: Legislators want agencies to be competent problem‐solvers, but they also want to keep agencies under control. How do EU legislators balance these two imperatives? We argue that agencies' political principals do not necessarily have to trade competence for control, and vice versa, but can draw on different institutional strategies – manipulating agencies' decision‐making and decision‐maker independence – to mitigate the competence‐control dilemma. Drawing on an original dataset on EU agencies' formal independence, we demonstrate empirically that principals customize EU agencies' independence to match the respective competence and control demands that come with specific agency tasks: regulation, authorization, implementation and information. The paper makes an important contribution to the literature on the drivers of the EU's agencification and to debates on the institutional design of regulatory institutions in the EU and beyond.

Synthetic data in cancer and cerebrovascular disease research: A novel approach to big data.

Synthetic datasets are artificially manufactured based on real health systems data but do not contain real patient information. We sought to validate the use of synthetic data in stroke and cancer research by conducting a comparison study of cancer patients with ischemic stroke to non-cancer patients with ischemic stroke. retrospective cohort study. We used synthetic data generated by MDClone and compared it to its original source data (i.e. real patient data from the Ottawa Hospital Data Warehouse). We compared key differences in demographics, treatment characteristics, length of stay, and costs between cancer patients with ischemic stroke and non-cancer patients with ischemic stroke. We used a binary, multivariable logistic regression model to identify risk factors for recurrent stroke in the cancer population. Using synthetic data, we found cancer patients with ischemic stroke had a lower prevalence of hypertension (52.0% in the cancer cohort vs 57.7% in the non-cancer cohort, p<0.0001), and a higher prevalence of chronic obstructive pulmonary disease (COPD: 8.5% vs 4.7%, p<0.0001), prior ischemic stroke (1.7% vs 0.1%, p<0.001), and prior venous thromboembolism (VTE: 8.2% vs 1.5%, p<0.0001). They also had a longer length of stay (8 days [IQR 3-16] vs 6 days [IQR 3-13], p = 0.011), and higher costs associated with their stroke encounters: $11,498 (IQR $4,440 -$20,668) in the cancer cohort vs $8,084 (IQR $3,947 -$16,706) in the non-cancer cohort (p = 0.0061). A multivariable logistic regression model identified 5 predictors for recurrent ischemic stroke in the cancer cohort using synthetic data; 3 of the same predictors identified using real patient data with similar effect measures. Summary statistics between synthetic and original datasets did not significantly differ, other than slight differences in the distributions of frequencies for numeric data. We demonstrated the utility of synthetic data in stroke and cancer research and provided key differences between cancer and non-cancer patients with ischemic stroke. Synthetic data is a powerful tool that can allow researchers to easily explore hypothesis generation, enable data sharing without privacy breaches, and ensure broad access to big data in a rapid, safe, and reliable fashion.

Defining artificial intelligence as a policy problem: A discourse network analysis from Germany

AbstractScholars agree that digital technologies such as artificial intelligence (AI) pose a political challenge. In this article, we study empirically how different actors in the German political system define AI as a policy problem. We use an original data set of 6421 statements by representatives of political parties, interest groups, scientific experts, and public officials in parliamentary debates, government consultations, and quality newspapers. Through Discourse Network Analysis and quantitative text analyses we show that most actors define AI as technology (innovation) policy and link it to government operations, international cooperation, and macroeconomics. Although they are present, consumer protection, labor, and education seem to be less important policy issues concerning AI. The results imply that the capacity of the national government to reduce problem definition uncertainty and to steer the political agenda is difficult and that most actors focus on technological innovation rather than civil rights‐related aspects.

LoHDP: Adaptive local differential privacy for high‐dimensional data publishing

SummaryThe increasing availability of high‐dimensional data collected from numerous users has led to the need for multi‐dimensional data publishing methods that protect individual privacy. In this paper, we investigate the use of local differential privacy for such purposes. Existing solutions calculate pairwise attribute marginals to construct probabilistic graphical models for generating attribute clusters. These models are then used to derive low‐dimensional marginals of these clusters, allowing for an approximation of the distribution of the original dataset and the generation of synthetic datasets. Existing solutions have limitations in computing the marginals of pairwise attributes and multi‐dimensional distribution on attribute clusters, as well as constructing relational dependency graphs that contain large clusters. To address these problems, we propose LoHDP, a high‐dimensional data publishing method composed of adaptive marginal computing and an effective attribute clustering method. The adaptive local marginal calculates any k‐dimensional marginals required in the algorithm. In particular, methods such as sampling‐based randomized response are used instead of privacy budget splits to perturb user data. The attribute clustering method measures the correlation between pairwise attributes using an effective method, reduces the search space during the construction of the dependency graph using high‐pass filtering technology, and realizes dimensionality reduction by combining sufficient triangulation operation. We demonstrate through extensive experiments on real datasets that our LoHDP method outperforms existing methods in terms of synthetic dataset quality.

Sentiment analysis based on BiLSTM with attention mechanism on Chinese comment with stickers

As the Internet is progressively becoming larger and more intricate, more and more users of various social media choose to post their comments to express their opinions and thinking on those platforms. Analyzing the emotions contained in user comments holds great business value, helping to accurately perceive user consumption habits and improve user service levels. However, the use of emoticons and stickers in comments has increased dramatically in recent years, which brings new challenges to text sentiment analysis based on natural language processing. In this paper, in order to alleviate the above problems, we propose a method for analyzing the sentiment of Chinese comments based on the attention mechanism and BiLSTM. Specifically, we partitioned the original dataset from the Weibo platform according to the number and type of emoticons in the comments. By analyzing the actual data, the specific features of emojis that affect the performance of sentiment analysis are identified, and corresponding explanations are given. In addition, a hypothesis is proposed to quantify the impact of emoticons on model effectiveness. All the results demonstrate the effectiveness of our proposed method.

Political Professionalization Beyond National Borders: An Analysis of Italian MPs in Overseas Constituencies

Among European countries, Italy is one of the relatively few cases to provide a quota of reserved parliamentary seats for non-resident citizens. Despite an increased scientific interest in the topic, the group of MPs elected in Overseas Constituencies remains overlooked in the available literature. The gap relates to factors such as their socio-biographical profile, precedent careers, parliamentary activity, the role played in their recruitment by the party abroad or at a national level, and their style of representation. In this vein, the article investigates the career profiles of Italian MPs elected in Overseas Constituencies from 2006 (the first elections with the introduction of citizen representatives living abroad) to 2022. Based on an original data set and through an analysis of their biographical and political characteristics, the article builds a typology of elected MPs abroad by cross-referencing two dimensions derived from the literature: the linkage with the host country and the presence of previous political and associational experiences. The analysis shows that different types of MPs have different career lengths and a different capacity to collect preference votes.

Reducing Social Stratification Bias in Referendum Participation: Evidence from the German Local Level

Political participation is socially distorted: Socio-economic resources affecting the probability of individuals becoming politically active question the principle of political equality. Such inequality in participation based on social stratification is well documented for elections, while research on inequality in referendum participation is still scarce. Based on the observation that such inequality varies between referendums, this paper explores referendum-specific contextual factors that may affect socially distorted referendum participation. We leverage information from an original dataset covering 1788 districts in 35 German municipalities for 68 local referendums held between 2000 and 2019. The results of our multilevel models indicate that concurrent first-order elections boost referendum turnout and decrease social stratification, while the closeness of the decision increases turnout but does not affect stratification bias. Moreover, we find a curvilinear relationship between turnout and stratification, with stratification only diminishing from participation levels above 40–50%.

Linking crash and breathalyzer data in Connecticut

Objectives To document the process of linking breathalyzer and motor vehicle crash (MVC) data for the State of Connecticut using a unique identifier in the place of personal and private information. Methods Deterministic linkage methodologies were utilized in Microsoft SQL Server to join 5,634 (of 6,650) breathalyzer records to corresponding MVC driver records for the period of January 1, 2017 to December 31, 2022. Differences between the linked and original datasets were documented by comparing the consistency of frequency and proportion distributions of key variables. Results Proportions of annual records, alcohol breath tests, and refusals were nearly unchanged when comparing linked and original breathalyzer data. When examining variables in the original MVC driver records, there were differences in the within-group proportions for sex and age, with an overrepresentation of males and drivers aged 26-to-40 years old. For crash and injury severity, the linked dataset had lower proportions of more severe injury records when compared to the original MVC data. Additionally, 1,007 breathalyzer records were not matched with an associated MVC record. Conclusions Linkage methodology is sound and produced quality matches. The use of a unique identifier provided a strong match qualifier in the absence of personal and private data. Changes in proportions for age, sex, crash and injury severity align with previous research. Potential missed matches may be attributed to several factors outside of the linkage process, including data discrepancies and varied reporting practices. Future studies will further explore these differences and incorporate additional toxicology data as part of a continued effort to fuze crash, citation, toxicology, and public health data. The end result will be a holistic, comprehensive, and multifaceted database for transportation research and education.

Design and development of artificial intelligence‐based application programming interface for early detection and diagnosis of colorectal cancer from wireless capsule endoscopy images

AbstractThe colorectal cancer (CRC) is gaining attention in the context of gastrointestinal tract diseases as it ranks third among the most prevalent type of cancer. The early diagnosis of the CRC can be done by periodic examination of the colon and rectum for innocuous tissue abnormality called polyp as it has the potential to evolve as malignant in future. The CRC diagnosis using wireless capsule endoscopy requires the dedicated commitment of the medical expert demanding significant time, focus and effort. The accuracy of manual analysis in identifying polyps is extensively reliant on the cognitive condition of the physician, thus emphasizing the requirement for automatic polyp identification. The artificial intelligence integrated computer‐aided diagnosis system could assist the clinician in better diagnosis, thereby reducing the miss‐rates of polyps. In our proposed study, we developed an application program interface to aid the clinician in automatic segmentation of the polyp and evaluation of its dimension by manual placement of four landmarks on the predicted polyp. The segmentation is performed by the proposed light weight Padded U‐Net for the effective polyp segmentation in the colorectal images. We trained and validated the Padded U‐Net with augmented images of Kvasir dataset and calculated the performance parameters. In order to facilitate image augmentation, a graphical user interface called Augment Tree was developed, which incorporates 92 augmentation techniques. The accuracy, recall, precision, IoU, F1‐score, loss achieved during validation of Padded U‐Net were 95.6%, 0.946%, 0.985%, 0.933%, 0.965% and 0.080% respectively. We demonstrated that accuracy was improved and loss was reduced when the model was trained with augmented images rather than only the limited original dataset images. On comparison of our Padded U‐net architecture with recently developed architectures, our model attained optimal performance in all the metrics except accuracy in which it attained marginal performance to the highest value.

Diplomatic Statements and the Strategic Use of Terrorism in Civil Wars

How does third-party diplomatic and material support affect rebel groups’ use of terrorism in civil wars? We argue via a game-theoretic model that diplomatic support prompts prospective shifts in rebel tactics, from civilian to military targets, in anticipation of material support, while material support alters the cost structure of attacks, leading to the same tactical shift. We empirically test the model’s implications using an original dataset of UN resolutions about countries in civil wars as well as a case study of South Africa. In support of our theory, we find that both diplomatic resolutions and material interventions in favor of the rebels are associated with rebel tactical shifts, leading to decreased reliance on violence against civilians. These findings demonstrate the value of modeling civilian and military targeting as substitutes rather than examining civilian targeting in isolation.

DiffusionFR: Species Recognition of Fish in Blurry Scenarios via Diffusion and Attention

Blurry scenarios, such as light reflections and water ripples, often affect the clarity and signal-to-noise ratio of fish images, posing significant challenges for traditional deep learning models in accurately recognizing fish species. Firstly, deep learning models rely on a large amount of labeled data. However, it is often difficult to label data in blurry scenarios. Secondly, existing deep learning models need to be more effective for the processing of bad, blurry, and otherwise inadequate images, which is an essential reason for their low recognition rate. A method based on the diffusion model and attention mechanism for fish image recognition in blurry scenarios, DiffusionFR, is proposed to solve these problems and improve the performance of species recognition of fish images in blurry scenarios. This paper presents the selection and application of this correcting technique. In the method, DiffusionFR, a two-stage diffusion network model, TSD, is designed to deblur bad, blurry, and otherwise inadequate fish scene pictures to restore clarity, and a learnable attention module, LAM, is intended to improve the accuracy of fish recognition. In addition, a new dataset of fish images in blurry scenarios, BlurryFish, was constructed and used to validate the effectiveness of DiffusionFR, combining bad, blurry, and otherwise inadequate images from the publicly available dataset Fish4Knowledge. The experimental results demonstrate that DiffusionFR achieves outstanding performance on various datasets. On the original dataset, DiffusionFR achieved the highest training accuracy of 97.55%, as well as a Top-1 accuracy test score of 92.02% and a Top-5 accuracy test score of 95.17%. Furthermore, on nine datasets with light reflection noise, the mean values of training accuracy reached a peak at 96.50%, while the mean values of the Top-1 accuracy test and Top-5 accuracy test were at their highest at 90.96% and 94.12%, respectively. Similarly, on three datasets with water ripple noise, the mean values of training accuracy reached a peak at 95.00%, while the mean values of the Top-1 accuracy test and Top-5 accuracy test were at their highest at 89.54% and 92.73%, respectively. These results demonstrate that the method showcases superior accuracy and enhanced robustness in handling original datasets and datasets with light reflection and water ripple noise.

Predicting the main pollen season of Broussonetia Papyrifera (paper mulberry) tree.

Paper mulberry pollen, declared a pest in several countries including Pakistan, can trigger severe allergies and cause asthma attacks. We aimed to develop an algorithm that could accurately predict high pollen days to underpin an alert system that would allow patients to take timely precautionary measures. We developed and validated two prediction models that take historical pollen and weather data as their input to predict the start date and peak date of the pollen season in Islamabad, the capital city of Pakistan. The first model is based on linear regression and the second one is based on phenological modelling. We tested our models on an original and comprehensive dataset from Islamabad. The mean absolute errors (MAEs) for the start day are 2.3 and 3.7 days for the linear and phenological models, respectively, while for the peak day, the MAEs are 3.3 and 4.0 days, respectively. These encouraging results could be used in a website or app to notify patients and healthcare providers to start preparing for the paper mulberry pollen season. Timely action could reduce the burden of symptoms, mitigate the risk of acute attacks and potentially prevent deaths due to acute pollen-induced allergy.

Susceptibility of AutoML mortality prediction algorithms to model drift caused by the COVID pandemic

BackgroundConcept drift and covariate shift lead to a degradation of machine learning (ML) models. The objective of our study was to characterize sudden data drift as caused by the COVID pandemic. Furthermore, we investigated the suitability of certain methods in model training to prevent model degradation caused by data drift.MethodsWe trained different ML models with the H2O AutoML method on a dataset comprising 102,666 cases of surgical patients collected in the years 2014–2019 to predict postoperative mortality using preoperatively available data. Models applied were Generalized Linear Model with regularization, Default Random Forest, Gradient Boosting Machine, eXtreme Gradient Boosting, Deep Learning and Stacked Ensembles comprising all base models. Further, we modified the original models by applying three different methods when training on the original pre-pandemic dataset: (1) we weighted older data weaker, (2) used only the most recent data for model training and (3) performed a z-transformation of the numerical input parameters. Afterwards, we tested model performance on a pre-pandemic and an in-pandemic data set not used in the training process, and analysed common features.ResultsThe models produced showed excellent areas under receiver-operating characteristic and acceptable precision-recall curves when tested on a dataset from January-March 2020, but significant degradation when tested on a dataset collected in the first wave of the COVID pandemic from April-May 2020. When comparing the probability distributions of the input parameters, significant differences between pre-pandemic and in-pandemic data were found. The endpoint of our models, in-hospital mortality after surgery, did not differ significantly between pre- and in-pandemic data and was about 1% in each case. However, the models varied considerably in the composition of their input parameters. None of our applied modifications prevented a loss of performance, although very different models emerged from it, using a large variety of parameters.ConclusionsOur results show that none of our tested easy-to-implement measures in model training can prevent deterioration in the case of sudden external events. Therefore, we conclude that, in the presence of concept drift and covariate shift, close monitoring and critical review of model predictions are necessary.

Optimizing road sign detection using the segment anything model for background pixel exclusion

Road sign recognition plays a great role in automatic driving. At present, for the task of road sign recognition, the structure related to CNN is mainly used to classify the detected road signs. However, due to the complexity of road conditions and weather, the detected images often contain complex background pixels, which affects the accuracy of classification. Due to the release of segment anything model, there is a powerful tool for eliminating background pixels quickly and conveniently. We optimize the data to retain only the road sign part of the images, and directly let the model train the image without back- ground pixels to realize the recognition task. This paper constructs two CNN classification models with the same structure through comparative experiments. One model uses the segmented data set for training, and the other model uses the original data set for training. The performance of the two models is evaluated and compared to verify the optimization effect brought by the segmented data set. It is found that segment anything model can accurately cut most of the images in the data set. And the performance of the model trained by the segmented data set is better than that of the model using the original data set. Moreover, the optimized model can achieve high accuracy in less training times.

Deep learning application to automated classification of recommendations made by hospital pharmacists during medication prescription review.

Recommendations to improve therapeutics are proposals made by pharmacists during the prescription review process to address suboptimal use of medicines. Recommendations are generated daily as text documents but are rarely reused beyond their primary use to alert prescribers and caregivers. If recommendation data were easier to summarize, they could be used retrospectively to improve safeguards for better prescribing. The objective of this work was to train a deep learning algorithm for automated recommendation classification to valorize the large amount of recommendation data. The study was conducted in a French university hospital, at which recommendation data were collected throughout 2017. Data from the first 6 months of 2017 were labeled by 2 pharmacists who assigned recommendations to 1 of the 29 possible classes of the French Society of Clinical Pharmacy classification. A deep neural network classifier was trained to predict the class of recommendations. In total, 27,699 labeled recommendations from the first half of 2017 were used to train and evaluate a classifier. The prediction accuracy calculated on a validation dataset was 78.0%. We also predicted classes for unlabeled recommendations collected during the second half of 2017. Of the 4,460 predictions reviewed, 67 required correction. When these additional labeled data were concatenated with the original dataset and the neural network was retrained, accuracy reached 81.0%. To facilitate analysis of recommendations, we have implemented an automated classification system using deep learning that achieves respectable performance. This tool can help to retrospectively highlight the clinical significance of daily medication reviews performed by hospital clinical pharmacists.

The costs and benefits of governing in a multi-level system

ABSTRACT In recent years the ‘cost of governing’ has significantly increased for some mainstream political parties. In a context of financial uncertainty, multiple crises and growing constraints exerted by global forces, being a ‘natural’ party of government is no longer regarded as an electoral advantage. This is particularly true for parties that have moved from a position of clear dominance within ruling coalitions to a more subordinate role. In this article, focusing on the Social Democratic Party of Germany (SPD) and using an original dataset, we aim to provide a more nuanced assessment of the effects of incumbency by examining regional electoral performance since 1990. It appears that sub-national incumbency can be beneficial in regional elections, especially when a party faces significant costs of governing at the national level. However, this advantage is only applicable if the party holds a leading position in the regional executive. On the contrary, being a junior coalition partner at both national and regional levels may further exacerbate electoral decline for the party.

Machine Learning Refinement of In Situ Images Acquired by Low Electron Dose LC-TEM.

We have studied a machine learning (ML) technique for refining images acquired during in situ observation using liquid-cell transmission electron microscopy. Our model is constructed using a U-Net architecture and a ResNet encoder. For training our ML model, we prepared an original image dataset that contained pairs of images of samples acquired with and without a solution present. The former images were used as noisy images, and the latter images were used as corresponding ground truth images. The number of pairs of image sets was 1,204, and the image sets included images acquired at several different magnifications and electron doses. The trained model converted a noisy image into a clear image. The time necessary for the conversion was on the order of 10 ms, and we applied the model to in situ observations using the software Gatan DigitalMicrograph (DM). Even if a nanoparticle was not visible in a view window in the DM software because of the low electron dose, it was visible in a successive refined image generated by our ML model.