Lecture notes in computer science, Год журнала: 2024, Номер unknown, С. 84 - 104
Опубликована: Янв. 1, 2024
Язык: Английский
Journal of Healthcare Informatics Research, Год журнала: 2023, Номер 8(1), С. 1 - 18
Опубликована: Сен. 20, 2023
Abstract Glioblastoma multiforme (GM) is a malignant tumor of the central nervous system considered to be highly aggressive and often carrying terrible survival prognosis. An accurate prognosis therefore pivotal for deciding good treatment plan patients. In this context, computational intelligence applied data electronic health records (EHRs) patients diagnosed with disease can useful predict patients’ time. study, we evaluated different machine learning models time in suffering from glioblastoma further investigated which features were most predictive We our methods three independent open datasets EHRs glioblastoma: Shieh dataset 84 patients, Berendsen 647 Lammer 60 Our prediction techniques obtained concordance index (C-index) = 0.583 dataset, C-index 0.776 0.64 as best results each dataset. Since original studies regarding analyzed here did not provide insights about clinical time, feature importance among these datasets. To end, then utilized Random Survival Forests, decision tree-based algorithm able model non-linear interaction between might better capture complex genetic status discoveries impact practice, aiding clinicians alike decide therapy suited their unique status.
Язык: Английский
Процитировано
7
Research Square (Research Square), Год журнала: 2023, Номер unknown
Опубликована: Июль 3, 2023
Abstract Continuous glucose monitoring (CGM) has transformed the care of diabetes mellitus patients. It is increasingly used to support nutritional management in various pathophysiologic conditions, including obesity and migraine, where avoiding postprandial hyperglycemia critical prevent hyperinsulinemia low levels after meal intake. However, current CGM devices have significant limitations, such as invasiveness, availability, high cost, limited shelf life, which must be overcome for broader use, specifically disease prevention management. Therefore, we collected over 15,500 interstitial measurements by with a large amount corresponding multimodal non-invasive wearable sensor data healthy cohorts machine learning (ML) approach discover their correlations accurately predict response (PPGR) real-time ten days, an average Root Mean Squared Error (RMSE) 18.49 ± 0.1 mg/dL Absolute Percentage (MAPE) 15.58 0.09%. demonstrated that ML could facilitate blood opened new opportunities efficient objective nutritional-based prevention.
Язык: Английский
Процитировано
4
PLoS Computational Biology, Год журнала: 2023, Номер 19(7), С. e1011272 - e1011272
Опубликована: Июль 20, 2023
Some scientific studies involve huge amounts of bioinformatics data that cannot be analyzed on personal computers usually employed by researchers for day-to-day activities but rather necessitate effective computational infrastructures can work in a distributed way. For this purpose, computing systems have become useful tools to analyze large and generate relevant results virtual environments, where software executed hours or even days without affecting the computer laptop researcher. Even if resources pivotal multiple laboratories, often students use them wrong ways, making mistakes cause underperform outcomes. In context, we present here ten quick tips usage Apache Spark analyses: simple guidelines that, taken into account, help users avoid common run their analyses smoothly. designed our recommendations beginners students, they should followed experts too. We think anyone make more efficiently ultimately better, reliable results.
Язык: Английский
Процитировано
4
medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown
Опубликована: Май 27, 2024
Abstract Importance Accurately predicting major bleeding events in non-valvular atrial fibrillation (AF) patients on direct oral anticoagulants (DOACs) is crucial for personalized treatment and improving patient outcomes, especially with emerging alternatives like left appendage closure devices. The devices reduce stroke risk comparably but significantly fewer non-procedural events. Objective To evaluate the performance of machine learning (ML) models clinically significant requiring hospitalization hemorrhagic AF DOACs compared to conventional scores (HAS-BLED, ORBIT, ATRIA) at index visit a cardiologist management. Design Prognostic modeling retrospective cohort study design using electronic health record (EHR) data, clinical follow-up one-, two-, five-years. Setting University Pittsburgh Medical Center (UPMC) system. Participants 24,468 aged ≥18 years treated DOACs, excluding those prior history bleeding, other indications warfarin or contraindicated DOACs. Exposure(s) DOAC therapy AF. Main Outcome(s) Measure(s) primary endpoint was within one year visit. incorporated demographic, clinical, laboratory variables available EHR Results Among patients, 553 (2.3%) had year, 829 (3.5%) two years, 1,292 (5.8%) five We evaluated multivariate logistic regression ML including random forest, classification trees, k-nearest neighbor, naive Bayes, extreme gradient boosting (XGBoost) which modestly outperformed HAS-BLED, ATRIA, ORBIT 1-year follow-up. best performing model (random forest) showed area under curve (AUC-ROC) 0.76 (0.70-0.81), G-Mean score 0.67, net reclassification 0.14 0.57 (0.50-0.63), HASBLED score, p-value difference <0.001. improved across time-points 2-year 5-years subgroup stroke. SHAP analysis identified novel factors measures from body mass index, cholesterol profile, insurance type beyond used scores. Conclusions Relevance Our findings demonstrate superior identify highlighting potential assessment
Язык: Английский
Процитировано
1
Lecture notes in computer science, Год журнала: 2024, Номер unknown, С. 84 - 104
Опубликована: Янв. 1, 2024
Язык: Английский
Процитировано
1