Performance Of Prediction Models Research Articles

Prediction of short-term progression of COVID-19 pneumonia based on chest CT artificial intelligence: during the Omicron epidemic

Background and purposeThe persistent progression of pneumonia is a critical determinant of adverse outcomes in patients afflicted with COVID-19. This study aimed to predict personalized COVID-19 pneumonia progression between the duration of two weeks and 1 month after admission by integrating radiological and clinical features.MethodsA retrospective analysis, approved by the Institutional Review Board, encompassed patients diagnosed with COVID-19 pneumonia between December 2022 and February 2023. The cohort was divided into training and validation groups in a 7:3 ratio. A trained multi-task U-Net network was deployed to segment COVID-19 pneumonia and lung regions in CT images, from which quantitative features were extracted. The eXtreme Gradient Boosting (XGBoost) algorithm was employed to construct a radiological model. A clinical model was constructed by LASSO method and stepwise regression analysis, followed by the subsequent construction of the combined model. Model performance was assessed using ROC and decision curve analysis (DCA), while Shapley’s Additive interpretation (SHAP) illustrated the importance of CT features.ResultsA total of 214 patients were recruited in our study. Four clinical characteristics and four CT features were identified as pivotal components for constructing the clinical and radiological models. The final four clinical characteristics were incorporated as well as the RS_radiological model to construct the combined prediction model. SHAP analysis revealed that CT score difference exerted the most significant influence on the predictive performance of the radiological model. The training group’s radiological, clinical, and combined models exhibited AUC values of 0.89, 0.72, and 0.92, respectively. Correspondingly, in the validation group, these values were observed to be 0.75, 0.72, and 0.81. The DCA curve showed that the combined model exhibited greater clinical utility than the clinical or radiological models.ConclusionOur novel combined model, fusing quantitative CT features with clinical characteristics, demonstrated effective prediction of COVID-19 pneumonia progression from 2 weeks to 1 month after admission. This comprehensive model can potentially serve as a valuable tool for clinicians to develop personalized treatment strategies and improve patient outcomes.

A stacking ensemble model for predicting soil organic carbon content based on visible and near-infrared spectroscopy

The content of soil organic carbon (SOC) plays an important role in maintaining ecosystem functions, protecting soil biodiversity, and understanding carbon cycling processes. The combination of visible and near-infrared spectroscopy (VIS–NIRS) and machine learning can achieve rapid prediction of SOC content. However, it is still relatively unknown how to integrate the characteristics of various machine learning models to improve the performance of SOC prediction models. In this study, a new model for predicting SOC content based on stacking ensemble learning was proposed by using VIS–NIRS. The prediction performances of six different models including Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Light Gradient Boosting Machine (LightGBM), PartialLeast-square (PLS) and Extreme Learning Machine (ELM) on SOC content under different spectral preprocessing methods were compared. The results indicated that SVR, XGBoost, and LightGBM models provide better prediction performance after first-order derivative preprocessing. After comparing the performance of various combinations of base models applied to the first layer of a stacking ensemble model, the results showed that both the combination of XGBoost, LightGBM, and SVR models and the combination of SVR, ELM, and LightGBM models achieve the best performance. The coefficient of determination (R2) of the stacking ensemble model on the test set reaches 0.84, which improves the accuracy of the model compared with the traditional single model. The stability of the stacking ensemble model was verified by applying it to datasets of different sizes, which can replace traditional machine learning models in predicting SOC content.

Towards a framework for reliable performance evaluation in defect prediction

Enhancing software reliability, dependability, and security requires effective identification and mitigation of defects during early development stages. Software defect prediction (SDP) models have emerged as valuable tools for this purpose. However, there is currently a lack of consensus in evaluating the predictive performance of newly proposed models, which hinders accurate measurement of progress and can lead to misleading conclusions. To tackle this challenge, we present MATTER (a fraMework towArd a consisTenT pErformance compaRison), which aims to provide reliable and consistent performance comparisons for SDP models. MATTER incorporates three key considerations. First, it establishes a global reference point, ONE (glObal baseliNe modEl), which possesses the 3S properties (Simplicity in implementation, Strong predictive ability, and Stable prediction performance), to serve as the baseline for evaluating other models. Second, it proposes using the SQA-effort-aligned threshold setting to ensure fair performance comparisons. Third, it advocates for consistent performance evaluation by adopting a set of core performance indicators that reflect the practical value of prediction models in achieving tangible progress. Through the application of MATTER to the same benchmark data sets, researchers and practitioners can obtain more accurate and meaningful insights into the performance of defect prediction models, thereby facilitating informed decision-making and improving software quality. When evaluating representative SDP models from recent years using MATTER, we surprisingly observed that: none of these models demonstrated a notable enhancement in prediction performance compared to the simple baseline model ONE. In future studies, we strongly recommend the adoption of MATTER to assess the actual usefulness of newly proposed models, promoting reliable scientific progress in defect prediction.

Prediction value study of breast cancer tumor infiltrating lymphocyte levels based on ultrasound imaging radiomics.

Construct models based on grayscale ultrasound and radiomics and compare the efficacy of different models in preoperatively predicting the level of tumor-infiltrating lymphocytes in breast cancer. This study retrospectively collected clinical data and preoperative ultrasound images from 185 breast cancer patients confirmed by surgical pathology. Patients were randomly divided into a training set (n=111) and a testing set (n=74) using a 6:4 ratio. Based on a 10% threshold for tumor-infiltrating lymphocytes (TIL) levels, patients were classified into low-level and high-level groups. Radiomic features were extracted and selected using the training set. The evaluation included assessing the relationship between TIL levels and both radiomic features and grayscale ultrasound features. Subsequently, grayscale ultrasound models, radiomic models, and nomograms combining radiomics score (Rad-score) and grayscale ultrasound features were established. The predictive performance of different models was evaluated through receiver operating characteristic (ROC) analysis. Calibration curves assessed the fit of the nomograms, and decision curve analysis (DCA) evaluated the clinical effectiveness of the models. Univariate analyses and multivariate logistic regression analyses revealed that indistinct margin (P<0.001, Odds Ratio [OR]=0.214, 95% Confidence Interval [CI]: 0.103-1.026), posterior acoustic enhancement (P=0.027, OR=2.585, 95% CI: 1.116-5.987), and ipsilateral axillary lymph node enlargement (P=0.001, OR=4.214, 95% CI: 1.798-9.875) were independent predictive factors for high levels of TIL in breast cancer. In comparison to grayscale ultrasound model (Training set: Area under curve [AUC] 0.795; Testing set: AUC 0.720) and radiomics model (Training set: AUC 0.803; Testing set: AUC 0.759), the nomogram demonstrated superior discriminative ability on both the training (AUC 0.884) and testing (AUC 0.820) datasets. Calibration curves indicated high consistency between the nomogram model's predicted probability of breast cancer TIL levels and the actual occurrence probability. DCA revealed that the radiomics model and the nomogram model achieved higher clinical net benefits compared to the grayscale ultrasound model. The nomogram based on preoperative ultrasound radiomics features exhibits robust predictive capacity for the non-invasive evaluation of breast cancer TIL levels, potentially providing a significant basis for individualized treatment decisions in breast cancer.

Rapid quantitative analysis of coal composition using laser-induced breakdown spectroscopy coupled with random forest algorithm.

Coal is the primary energy source in China, widely used in energy production, industrial processes, and chemical engineering. Due to the complexity and diversity of coal quality, there is an urgent need for new technologies to achieve rapid and accurate detection and analysis of coal, aiming to improve coal resource utilization and reduce pollutant emissions. This study proposes a rapid quantitative analysis of coal using laser-induced breakdown spectroscopy combined with the random forest algorithm. Firstly, a Q-switched Nd: YAG laser at 1064nm was employed to ablate coal samples, generating plasma, and spectral data were collected using a spectrometer. Secondly, the study explores the impact of different parameters in the preprocessing method (wavelet transform) on the predictive performance of the random forest model. It identifies elements related to coal ash content and calorific value along with their spectral information. Subsequently, to further validate the predictive performance of the model, a comparison is made with models established using support vector machine, artificial neural network, and partial least squares. Finally, under optimal parameters for spectral information preprocessing (wavelet transform with Db4 as the base function and 3 decomposition levels), a model combining wavelet transform with Random Forest is established to predict and analyze the ash content and calorific value of coal. The results demonstrate that the Wavelet Transform-Random Forest model exhibits excellent predictive performance (coal ash content: R2 = 0.9470, RMSECV = 4.8594, RMSEP = 4.8450; coal calorific value: R2 = 0.9485, RMSECV = 1.5996, RMSEP = 1.5949). Therefore, laser-induced breakdown spectroscopy combined with the random forest algorithm is an effective method for rapid and accurate detection and analysis of coal. The predicted coal composition values show high accuracy, providing insights and methods for coal composition monitoring and analysis.

Predicting treatment response in multicenter non-small cell lung cancer patients based on federated learning

BackgroundMulticenter non-small cell lung cancer (NSCLC) patient data is information-rich. However, its direct integration becomes exceptionally challenging due to constraints involving different healthcare organizations and regulations. Traditional centralized machine learning methods require centralizing these sensitive medical data for training, posing risks of patient privacy leakage and data security issues. In this context, federated learning (FL) has attracted much attention as a distributed machine learning framework. It effectively addresses this contradiction by preserving data locally, conducting local model training, and aggregating model parameters. This approach enables the utilization of multicenter data with maximum benefit while ensuring privacy safeguards. Based on pre-radiotherapy planning target volume images of NSCLC patients, a multicenter treatment response prediction model is designed by FL for predicting the probability of remission of NSCLC patients. This approach ensures medical data privacy, high prediction accuracy and computing efficiency, offering valuable insights for clinical decision-making.MethodsWe retrospectively collected CT images from 245 NSCLC patients undergoing chemotherapy and radiotherapy (CRT) in four Chinese hospitals. In a simulation environment, we compared the performance of the centralized deep learning (DL) model with that of the FL model using data from two sites. Additionally, due to the unavailability of data from one hospital, we established a real-world FL model using data from three sites. Assessments were conducted using measures such as accuracy, receiver operating characteristic curve, and confusion matrices.ResultsThe model’s prediction performance obtained using FL methods outperforms that of traditional centralized learning methods. In the comparative experiment, the DL model achieves an AUC of 0.718/0.695, while the FL model demonstrates an AUC of 0.725/0.689, with real-world FL model achieving an AUC of 0.698/0.672.ConclusionsWe demonstrate that the performance of a FL predictive model, developed by combining convolutional neural networks (CNNs) with data from multiple medical centers, is comparable to that of a traditional DL model obtained through centralized training. It can efficiently predict CRT treatment response in NSCLC patients while preserving privacy.

Optimizing clinico-genomic disease prediction across ancestries: a machine learning strategy with Pareto improvement

BackgroundAccurate prediction of an individual’s predisposition to diseases is vital for preventive medicine and early intervention. Various statistical and machine learning models have been developed for disease prediction using clinico-genomic data. However, the accuracy of clinico-genomic prediction of diseases may vary significantly across ancestry groups due to their unequal representation in clinical genomic datasets.MethodsWe introduced a deep transfer learning approach to improve the performance of clinico-genomic prediction models for data-disadvantaged ancestry groups. We conducted machine learning experiments on multi-ancestral genomic datasets of lung cancer, prostate cancer, and Alzheimer’s disease, as well as on synthetic datasets with built-in data inequality and distribution shifts across ancestry groups.ResultsDeep transfer learning significantly improved disease prediction accuracy for data-disadvantaged populations in our multi-ancestral machine learning experiments. In contrast, transfer learning based on linear frameworks did not achieve comparable improvements for these data-disadvantaged populations.ConclusionsThis study shows that deep transfer learning can enhance fairness in multi-ancestral machine learning by improving prediction accuracy for data-disadvantaged populations without compromising prediction accuracy for other populations, thus providing a Pareto improvement towards equitable clinico-genomic prediction of diseases.

Multifractal Characteristics and Displacement Prediction of Deformation on Tunnel Portal Slope of Shallow Buried Tunnel Adjacent to Important Structures

The tunnel portal section is often in extremely weak and fragmented strata, and the deformation of the portal side and slope will affect the stability of the surrounding rock and the tunnel-supporting structure. However, the deformation characteristics and displacement development patterns of slopes in the tunnel portal section are not clear. In this paper, the multifractal characteristics and displacement prediction of the deformation sequence of the tunnel portal slope at of a weak and water-rich shallow buried tunnel adjacent to an important structure are studied in depth. Combined with the deformation characteristics of the tunnel portal slope, a suitable slope monitoring and measurement scheme is designed to analyze the deformation pattern of the tunnel portal slope. Based on the multifractal detrended fluctuation analysis (MF-DFA) method, the multifractal characteristics of the deformation monitoring sequences at each monitoring point of the tunnel portal slope are analyzed. The multifractal characteristics of displacement sequences at different monitoring points of the tunnel portal slope are consistent with the actual monitoring results. Furthermore, the Long Short-Term Memory (LSTM) model is optimized using the Particle Swarm Optimization (PSO) algorithm to predict the deformation of the tunnel portal slope. The results show that the maximum mean square error (MSE) of the horizontal displacement test set prediction results is 0.142, and the coefficient of determination (R2) is higher than 91%. The maximum value of MSE for vertical displacement test set prediction is 0.069, and the R2 are higher than 91%. The study shows that the performance of the PSO-LSTM prediction model can meet the requirements for predicting the displacement of the tunnel portal slope. Based on the MF-DFA method and PSO-LSTM prediction model, the fluctuation characteristics of the displacement value of the tunnel portal section can be accurately identified and the displacement development pattern can be effectively predicted. The conclusions of the study are of great practical significance for the safe construction of the tunnel portal section.

Enabling active visitor management: local, short-term occupancy prediction at a touristic point of interest

AbstractAfter the temporary shock of the Covid-19 pandemic, the rapid recovery and resumed growth of the tourism sectors accelerates unsustainable tourism, resulting in local (over-)crowding, environmental damage, increased emissions, and diminished tourism acceptance. Addressing these challenges requires an active visitor management system at points of interest (POI), which requires local and timely POI-specific occupancy predictions to predict and mitigate crowding. Therefore, we present a new approach to measure visitor movement at an open-spaced, and freely accessible POI and evaluate the prediction performance of multiple occupancy and visitor count machine learning prediction models. We analyze multiple case combinations regarding spatial granularity, time granularity, and prediction time horizons. With an analysis of the SHAP values we determine the influence of the most important features on the prediction and extract transferable knowledge for similar regions lacking visitor movement data. The results underline that POI-specific prediction is achievable with a moderate relation for occupancy prediction and a strong relation for visitor count prediction. Across all cases, XGBoost and Random Forest outperform other models, with prediction accuracy increasing as the prediction time horizon shortens. For effective active visitor management, combining multiple models with different spatial aggregations and prediction time horizons provides the best information basis to identify appropriate steering measures. This innovative application of digital technologies facilitates information exchange between destination management organizations and tourists, promoting sustainable destination development and enhancing tourism experience.

AI-based approach to enable proactive identification of early lung cancer: A retrospective population health study and economic model.

10536 Background: Approximately 48,500 new cases of lung cancer are diagnosed each year in the UK with a five-year survival rate which lags behind other European countries. Most patients are diagnosed at stage III or IV disease, which has a significantly worse prognosis. The integration of real-world data and AI techniques presents a promising avenue for improvement, using novel solutions not achievable through traditional care models. This study explores proactive patient identification at scale, aligning with the "left shift" paradigm in population health. The approach uses innovative technologies and novel data sources to identify patients with early-stage disease, enhancing both clinical and economic outcomes. Importantly, these interventions should be feasible on a large scale without overburdening already stretched clinical services. Methods: A retrospective study of real-world data for an urban population with high socioeconomic disadvantage and health needs was undertaken. Two endpoints were determined: detection potential and health economic impact. For detection potential (DP), patients aged 40 and above who underwent chest X-rays (CXR) at a large academic medical centre between 2016-2022 were included. Natural Language Processing (NLP) analysed unstructured free-text data from Electronic Health Records, combined with ICD-10 codes, for predictive model (PM) development using Machine Learning (ML) classification techniques. The PM's performance was assessed using the area under the receiver operating characteristic curve (AUCROC). Health economic impact was evaluated through a theoretical extrapolation of DP. Assuming a 9% left shift from late (stage III or IV) to early (stage I or II) disease, a health economic model (HEM) compared the impact of the current care model to a potential future state facilitated by the digital medicine care model (DMCM). Results: 75,342 patients were included, with 755 lung cancer diagnoses occurring within 6 months of CXR. The PM achieved a peak AUROC of 0.75, exceeding the performance of currently used risk prediction models (Q Cancer Risk, Risk Assessment Tools). A 9% left shift equated to 4.5 years mean increase in survival. The DMCM was associated with a net healthcare cost benefit of £6k per patient diagnosed with lung cancer, improved quality-adjusted life years of £21M for the year 1 cohort over 10 years and increased economic productivity (from additional years in workforce) of £2.4M. Conclusions: The study demonstrates the practical utility of AI-supported early lung cancer detection at population scale. Real-world evidence is needed to prospectively validate the model as cost-effective, scalable, and efficient. Future work will seek to improve the model's performance against a broader at-risk patient population and explore its applicability to other late-presenting tumours.

Clinical T1/2 renal cell carcinoma: multiparametric dynamic contrast-enhanced MRI features-based model for the prediction of individual adverse pathology

BackgroundThe detection of renal cell carcinoma (RCC) has been rising due to the enhanced utilization of cross-sectional imaging and incidentally discovered lesions with adverse pathology demonstrate potential for metastasis. The purpose of our study was to determine the clinical and multiparametric dynamic contrast-enhanced magnetic resonance imaging (CEMRI) associated independent predictors of adverse pathology for cT1/2 RCC and develop the predictive model.MethodsWe recruited 105 cT1/2 RCC patients between 2018 and 2022, all of whom underwent preoperative CEMRI and had complete clinicopathological data. Adverse pathology was defined as RCC patients with nuclear grade III-IV; pT3a upstage; type II papillary RCC, collecting duct or renal medullary carcinoma, unclassified RCC; sarcomatoid/rhabdoid features. The qualitative and quantitative CEMRI parameters were independently reviewed by two radiologists. Univariate and multivariate binary logistic regression analyses were utilized to determine the independent predictors of adverse pathology for cT1/2 RCC and construct the predictive model. The receiver operating characteristic (ROC) curve, confusion matrix, calibration plot, and decision curve analysis (DCA) were conducted to compare the diagnostic performance of different predictive models. The individual risk scores and linear predicted probabilities were calculated for risk stratification, and the Kaplan–Meier curve and log-rank tests were used for survival analysis.ResultsOverall, 45 patients were pathologically confirmed as RCC with adverse pathology. Clinical characteristics, including gender, and CEMRI parameters, including RENAL score, tumor margin irregularity, necrosis, and tumor apparent diffusion coefficient (ADC) value were identified as independent predictors of adverse pathology for cT1/2 RCC. The clinical-CEMRI predictive model yielded an area under the curve (AUC) of the ROC curve of 0.907, which outperformed the clinical model or CEMRI signature model alone. Good calibration, better clinical usefulness, excellent risk stratification ability of adverse pathology and prognosis were also achieved for the clinical-CEMRI predictive model.ConclusionsThe proposed clinical-CEMRI predictive model offers the potential for preoperative prediction of adverse pathology for cT1/2 RCC. With the ability to forecast adverse pathology, the predictive model could significantly benefit patients and clinicians alike by providing enhanced guidance for treatment planning and decision-making.

Feasibility Study and Optimization of Limestone Calcined Clay Composites for Compressive Strength using Multi-Layered Explainable Artificial Intelligence Models

Interpretable white box machine learning models to predict the compressive strength of Limestone Calcined Clay (LC3) cementitious composite have been developed and analyzed. For this, a large dataset consisting of more than 4000 numerical instances has been collected from the literature, which has been tuned and normalized. Mainly based on the proportion of the chemical composition of its main ingredients – Limestone (0-31 by %mass of LC3), Calcined Clay (0-70 by %mass of LC3), and Cement (30-100 by %mass of LC3) with a water-to-binder ratio ranging from 0.40 to 0.60, predictive models have been developed for the 28 days compressive strength (fc) of LC3. The performance of various predictive models is then compared across the dataset, and suggestions are made for their reliability and explainability. The models are further interpretated using multi-layered explanations invoking model-agnostic explainable artificial intelligence (XAI) technique. Optimization and interaction plots have also been generated and shown as three-dimensional (3D) visualization. Percent of calcined clay, cement and the water to binder ratio were found to be the most influencing parameters for the compressive strength of LC3. Results of this study indicate that the optimized blends of LC3 can be expressed with up to 50% replacement of OPC without any significant compromise in fc, contributing to the development of sustainable cementitious composites. The major advantage of using this framework lies in the model agnostic approach, so different predictive models can be compared in a same manner and thus, this approach is applicable to interpretable forecasting models for binary, ternary, quaternary sustainable cementitious composites.

Explainable machine learning predicts survival of retroperitoneal liposarcoma: A study based on the SEER database and external validation in China

AbstractObjectiveWe have developed explainable machine learning models to predict the overall survival (OS) of retroperitoneal liposarcoma (RLPS) patients. This approach aims to enhance the explainability and transparency of our modeling results.MethodsWe collected clinicopathological information of RLPS patients from The Surveillance, Epidemiology, and End Results (SEER) database and allocated them into training and validation sets with a 7:3 ratio. Simultaneously, we obtained an external validation cohort from The First Affiliated Hospital of Naval Medical University (Shanghai, China). We performed LASSO regression and multivariate Cox proportional hazards analysis to identify relevant risk factors, which were then combined to develop six machine learning (ML) models: Cox proportional hazards model (Coxph), random survival forest (RSF), ranger, gradient boosting with component‐wise linear models (GBM), decision trees, and boosting trees. The predictive performance of these ML models was evaluated using the concordance index (C‐index), the integrated cumulative/dynamic area under the curve (AUC), and the integrated Brier score, as well as the Cox–Snell residual plot. We also used time‐dependent variable importance, analysis of partial dependence survival plots, and the generation of aggregated survival SHapley Additive exPlanations (SurvSHAP) plots to provide a global explanation of the optimal model. Additionally, SurvSHAP (t) and survival local interpretable model‐agnostic explanations (SurvLIME) plots were used to provide a local explanation of the optimal model.ResultsThe final ML models are consisted of six factors: patient's age, gender, marital status, surgical history, as well as tumor's histopathological classification, histological grade, and SEER stage. Our prognostic model exhibits significant discriminative ability, particularly with the ranger model performing optimally. In the training set, validation set, and external validation set, the AUC for 1, 3, and 5 year OS are all above 0.83, and the integrated Brier scores are consistently below 0.15. The explainability analysis of the ranger model also indicates that histological grade, histopathological classification, and age are the most influential factors in predicting OS.ConclusionsThe ranger ML prognostic model exhibits optimal performance and can be utilized to predict the OS of RLPS patients, offering valuable and crucial references for clinical physicians to make informed decisions in advance.

Independent validation of 30-day emergency department visit after initial chemotherapy A-PROACCT model.

e23243 Background: This study assessed model performance and, more importantly, the clinical usefulness of the PROACCT model created by Grant et al. (JAMA, 2019), refined as A-PROACCT by Stein et al. (JOP, 2023) with data from Orlando Health. The predictive performance of models anticipating acute care events within thirty days (ACE30) of initial chemotherapy administration (ICA) were evaluated. Methods: Stein et al. models classified high-risk patients as those whose model output risk score was at or above the 90th percentile. They studied 5153 ICAs between July and November of 2021, of which 559 (10.8%) had an ACE30. Their first model used drug category, age group, ED visits and hospitalizations in the prior year, and cancer type, and had sensitivity (Sen), specificity, (Spec), C-Statistic (C-Stat), and positive predictive value (PPV) of 0.28, 0.93, 0.79 and 0.18 respectively. Their second model added insurance category, number of anti-cancer agents, race, and BMI category and reported Sen, Spec, C-Stat, and PPV of 0.30, 0.94, 0.80 and 0.23 respectively. Of the high-risk patients, 33.6% and 34.1% had an ACE30 respectively. Of the low-risk patients, 8.3% and 8.1% had an ACE30 respectively. Our data were propagated through the two models. Results: Between February of 2012 and April 2021, Orlando Health had 15,483 distinct patient visits with ICAs. Overall, 2046 (13.2%) had an ACE30. Model one had Sen, Spec, C-Stat, and PPV of 0.19, 0.92, 0.65 and 0.27 respectively. Model two had Sen, Spec, C-Stat, and PPV of 0.19, 0.91, 0.65 and 0.25 respectively. Both models had an accuracy of 0.82. This demonstrated our data fit the Stein model. Of the 1460 ICAs identified as high-risk by model one, 387 (26.5%) had an ACE30, and of the 14,023 ICAs identified as low-risk, 1659 (11.8%) had an ACE30. Of the 1549 ICAs identified as high-risk by model two, 393 (25.4%) had an ACE30, and of the 13,934 ICAs identified as low-risk, 1653 (11.9%) had an ACE30. The difference of ACE30 rates between high risk and low risk groups were statistically significant (p &lt; 0.0001) in both models. This demonstrated both models were clinically useful in identifying chemotherapy patients at high risk of acute care events. Conclusions: These models allowed for clinically relevant stratification of patients into meaningful high-risk and low-risk groups (26.5% vs 11.8% with our data) for risk of acute care events within 30 days of initial chemotherapy administration. Nevertheless, only 26.5% of these high-risk patients actually had an acute care event within thirty days of initial chemotherapy administration. Further research is needed to improve the precision of identifying high-risk patients, by potentially adding more clinical variables, using alternative sampling methods, and employing nonlinear and more complex machine learning methodologies.

A genetic algorithm-based feature selection approach for diabetes prediction

&lt;p&gt;Genetic algorithms have emerged as a powerful optimization technique for feature selection due to their ability to search through a vast feature space efficiently. This study discusses the importance of feature selection for prediction in healthcare and prominently focuses on diabetes mellitus. Feature selection is essential for improving the performance of prediction models, by finding significant features and removing unnecessary among them. The study aims to identify the most informative subset of features. Diabetes is a chronic metabolic disorder that poses significant health challenges worldwide. For the experiment, two datasets related to diabetes were downloaded from Kaggle and the results of both (datasets) with and without feature selection using the genetic algorithm were compared. Machine learning classifiers and genetic algorithms were combined to increase the precision of diabetes risk prediction. In the preprocessing phase, feature selection, machine learning classifiers, and performance metrics methods were applied to make this study feasible. The results of the experiment showed that genetic algorithm + logistic regression i.e., 80% (accuracy) works better for PIMA diabetes, and for Germany diabetes dataset genetic algorithm + random forest and genetic algorithm + K-Nearest Neighbor i.e., 98.5% performed better than other chosen classifiers. The researchers can better comprehend the importance of feature selection in healthcare through this study.&lt;/p&gt;

MPSA: Multi-Position Supervised Soft Attention-based convolutional neural network for histopathological image classification

In recent years, significant achievements have been made in the field of histopathological image analysis using convolutional neural networks (CNNs). However, existing CNNs fail to fully capture the important local structures and regional information in histopathological images due to the complex tissue structures and variable pathological features present in these images. They often treat all regions equally, which further exacerbates the challenge of accurately analyzing such images. Current network model can’t extract deep layer features efficiently without guiding. To alleviate this problem, we propose a novel network model called Multi-Position Supervised Soft Attention (MPSA). MPSA adds regions of interest (RoI) labels at multiple feature layers for deep supervision, and then uses the supervised layers as soft attention to guide the learning of the classification network, enabling the network to accurately extract features of the lesion target. Additionally, we design a Multi-level Attention Feature Enhancement Module (MAFEM), which combines multiple levels of attention mechanisms to enhance the performance of the convolutional neural network in histopathological image classification. MAFEM includes spatial attention, soft attention of the main branch, and our proposed soft attention for multi-branch feature fusion. The proposed soft attention for multi-branch feature fusion aims to enhance the predictive performance of the classification model by activating relevant neurons in the diagnostic area in a highly activated state, while effectively avoiding noise activation. This innovative approach ensures that the model can focus on the most pertinent information, leading to improved classification accuracy. We conducted classification experiments on the liver cancer histopathological images dataset and the results showed that our method achieved a classification accuracy of 95.79%, indicating that it is very effective in the analysis of liver histopathological images. Our proposed network architecture has also demonstrated good generalization ability in other medical datasets, achieving a classification accuracy of 84.41% on the ultrasound carotid plaque dataset.

Tracking COPD exacerbation patterns and forecasting readmission risks utilizing electronic medical records

IntroductionAccurate evaluation of exacerbation frequency is an essential part of COPD assessment. But relying on just the prior-year exacerbation history may not capture the full picture of risk given the inherent year-to-year fluctuations in exacerbation rates. This study aimed to evaluate the predictive performance of models incorporating the 3-year exacerbation history based on electronic medical record. Materials and MethodsThis retrospective cohort study included 86,501 COPD hospitalized patients in Beijing from 2008 to 2014. The annual frequency of COPD exacerbation hospitalizations over a 3-year period after the index hospitalization was calculated, with patients segmented into seven distinct exacerbation trajectory groups. Logistic regression was used to evaluate the predictive capability of the 3-year exacerbation history for exacerbation readmission in the fourth year. Predictors included age, sex, comorbidities, and exacerbation hospitalization in previous 1–3 years. Model performance was evaluated using area under the receiver operating characteristic curve (AUC). ResultsOf the studied patients, 56.5% were men, and the mean age (SD) was 73.8 (10.3) years. The overall readmission rate for COPD exacerbation was 0.31 per person-year, with only 3.8% of patients persistently readmitted over three consecutive years. The 3-year trajectory of exacerbation frequency was associated with exacerbation risk in the fourth year. Compared to just the prior year, the inclusion of a 3-year exacerbation hospitalization history notably improved prediction accuracy, with AUC elevating from 0.731 (0.724–0.739) to 0.786 (0.779–0.792). ConclusionThese results unveil the fluctuating nature of COPD exacerbation hospitalization frequency across years and demonstrate that integrating a more comprehensive 3-year exacerbation history significantly refines the prediction model for future risk, thus providing a more nuanced and actionable insight for clinical care.

A deep learning approach for acute liver failure prediction with combined fully connected and convolutional neural networks.

Acute Liver Failure (ALF) is a critical medical condition with rapid development, often caused by viral infections, hepatotoxic drug abuse, or other severe liver diseases. Timely and accurate prediction of ALF occurrence is clinically crucial. However, predicting ALF poses challenges due to the diverse physiological differences among patients and the dynamic nature of the disease. This study introduces a deep learning approach that combines fully connected and convolutional neural networks for effective ALF prediction. The goal is to overcome limitations of traditional machine learning methods and enhance predictive model performance and generalization. The proposed model integrates a fully connected neural network for handling basic patient features and a convolutional neural network dedicated to capturing temporal patterns in patient data. The combination allows automatic learning of complex patterns and abstract features present in highly dynamic medical data associated with ALF. The model's effectiveness is demonstrated through comprehensive experiments and performance evaluations. It outperforms traditional machine learning methods, achieving 94.8% accuracy and superior generalization capabilities. The study highlights the potential of deep learning in ALF prediction, emphasizing the importance of considering individualized medical factors. Future research should focus on improving model robustness, addressing imbalanced data, and further exploring personalized features for enhanced predictive accuracy in real-world clinical scenarios.

Improving electricity demand forecasting accuracy: a novel grey-genetic programming approach using GMC(1,N) and residual sign estimation

PurposeThis paper addresses the challenges associated with forecasting electricity consumption using limited data without making prior assumptions on normality. The study aims to enhance the predictive performance of grey models by proposing a novel grey multivariate convolution model incorporating residual modification and residual genetic programming sign estimation.Design/methodology/approachThe research begins by constructing a novel grey multivariate convolution model and demonstrates the utilization of genetic programming to enhance prediction accuracy by exploiting the signs of forecast residuals. Various statistical criteria are employed to assess the predictive performance of the proposed model. The validation process involves applying the model to real datasets spanning from 2001 to 2019 for forecasting annual electricity consumption in Cameroon.FindingsThe novel hybrid model outperforms both grey and non-grey models in forecasting annual electricity consumption. The model's performance is evaluated using MAE, MSD, RMSE, and R2, yielding values of 0.014, 101.01, 10.05, and 99% respectively. Results from validation cases and real-world scenarios demonstrate the feasibility and effectiveness of the proposed model. The combination of genetic programming and grey convolution model offers a significant improvement over competing models. Notably, the dynamic adaptability of genetic programming enhances the model's accuracy by mimicking expert systems' knowledge and decision-making, allowing for the identification of subtle changes in electricity demand patterns.Originality/valueThis paper introduces a novel grey multivariate convolution model that incorporates residual modification and genetic programming sign estimation. The application of genetic programming to enhance prediction accuracy by leveraging forecast residuals represents a unique approach. The study showcases the superiority of the proposed model over existing grey and non-grey models, emphasizing its adaptability and expert-like ability to learn and refine forecasting rules dynamically. The potential extension of the model to other forecasting fields is also highlighted, indicating its versatility and applicability beyond electricity consumption prediction in Cameroon.

Artificial neural network modelling of natural gas dehydration process

A multi-input–multi-output artificial neuron network (MIMO-ANN) model has been developed for process monitoring and improvement on a natural gas glycol dehydration process. The MIMO-ANN model was based on a steady-state process simulation model constructed in commercial software Aspen HYSYS. A set of training data was generated with the converged simulation model for the training of the MIMO-ANN model in Python. The process input of the model includes lean glycol recirculation rate and purity along with wet gas inlet pressure. On the other hand, the process output considered includes dehydrated gas water and aromatics content, dehydrated gas hydrate formation temperature and water dew point, stripping gas flow rate as well as reboiler duty. The overall mean squared error (MSE) of the MIMO-ANN model was calculated as 1.79. The best-fit line that highly overlaps with a 45° diagonal line is constructed with a correlation coefficient (R2) score of more than 0.999 for all studied process output testing datasets. The mean absolute percentage error (MAPE) of the predicted outputs is generally less than 1%, except for dehydrated gas water dew point (16.63%) and stripping gas flow rate (11.55%), due to error predictions attempted on pseudo-zero values. This successfully displays an exceptional predictive performance of the MIMO-ANN model developed in this work which can be further deployed as an online dashboard for real-time monitoring tool.