Word2vec Word Embedding-Based Artificial Intelligence Model in the Triage of Patients with Suspected Diagnosis of Major Ischemic Stroke: A Feasibility Study

Desai, Antonio; Zumbo, Aurora; Giordano, Mauro; Morandini, Pierandrea; Laino, Maria Elena; Azzolini, Elena; Fabbri, Andrea; Marcheselli, Simona; Lucifero, Alice Giotta; Luzzi, Sabino; Voza, Antonio

doi:10.3390/ijerph192215295

Cited by 7 publications

(4 citation statements)

References 35 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hypothesis generated by the enhanced reading can inform surveys or be part of the design of clinical trials. For instance, the ETHOS survey [18, 19] used this tool to inform the panel about the semantic relevance of a list of preoperative features and postoperative outcomes. The panel expressed its opinion on approving or rejecting the proposed drivers after reviewing the quantitative information received.…”

Section: Discussionmentioning

confidence: 99%

Natural Language Processing to Extract Meaningful Information from a Corpus of Written Knowledge in Breast Cancer: Transforming Books into Data

Catanuto¹,

Rocco²,

Balafa³

et al. 2023

Breast Care

View full text Add to dashboard Cite

Introduction: Books and papers are the most relevant source of theoretical knowledge for medical education. New technologies of artificial intelligence can be designed to assist in selected educational tasks, such as reading a corpus made up of multiple documents and extracting relevant information in a quantitative way. Methods: Thirty experts were selected transparently using an online public call on the website of the sponsor organization and on its social media. Six books edited or co-edited by members of this panel containing a general knowledge of breast cancer or specific surgical knowledge have been acquired. This collection was used by a team of computer scientists to train an artificial neural network based on a technique called Word2Vec. Results: The corpus of six books contained about 2.2 billion words for 300d vectors. A few tests were performed. We evaluated cosine similarity between different words. Discussion: This work represents an initial attempt to derive formal information from textual corpus. It can be used to perform an augmented reading of the corpus of knowledge available in books and papers as part of a discipline. This can generate new hypothesis and provide an actual estimate of their association within the expert opinions. Word embedding can also be a good tool when used in accruing narrative information from clinical notes, reports, etc., and produce prediction about outcomes. More work is expected in this promising field to generate “real-world evidence.”

show abstract

Section: Discussionmentioning

confidence: 99%

Natural Language Processing to Extract Meaningful Information from a Corpus of Written Knowledge in Breast Cancer: Transforming Books into Data

Catanuto¹,

Rocco²,

Balafa³

et al. 2023

Breast Care

View full text Add to dashboard Cite

show abstract

“…Word2Vec technology is a crucial technique in the field of natural language processing. − It transforms vocabulary into computable low-dimensional vector representations, solving the issues of high dimensionality and inaccurate semantic similarity in traditional methods. In protein research, understanding the semantic information and functions of the protein sequence language is vital for revealing biological processes and disease mechanisms.…”

Section: Introductionmentioning

confidence: 99%

MIBPred: Ensemble Learning-Based Metal Ion-Binding Protein Classifier

Zhang,

Liu,

et al. 2024

ACS Omega

View full text Add to dashboard Cite

In biological organisms, metal ion-binding proteins participate in numerous metabolic activities and are closely associated with various diseases. To accurately predict whether a protein binds to metal ions and the type of metal ion-binding protein, this study proposed a classifier named MIBPred. The classifier incorporated advanced Word2Vec technology from the field of natural language processing to extract semantic features of the protein sequence language and combined them with positionspecific score matrix (PSSM) features. Furthermore, an ensemble learning model was employed for the metal ion-binding protein classification task. In the model, we independently trained XGBoost, LightGBM, and CatBoost algorithms and integrated the output results through an SVM voting mechanism. This innovative combination has led to a significant breakthrough in the predictive performance of our model. As a result, we achieved accuracies of 95.13% and 85.19%, respectively, in predicting metal ion-binding proteins and their types. Our research not only confirms the effectiveness of Word2Vec technology in extracting semantic information from protein sequences but also highlights the outstanding performance of the MIBPred classifier in the problem of metal ion-binding protein types. This study provides a reliable tool and method for the in-depth exploration of the structure and function of metal ion-binding proteins.

show abstract

“…With the aging of the global population, an increase in the proportion of elderly patients presenting with traumatic brain injury (TBI), defined as a disturbance in the functioning of the brain or evidence of brain pathology that results from an external physical force, is expected [ 1 , 2 , 3 ]. Indeed, in recent years, instances of neurotrauma in the elderly have been increasing [ 4 , 5 ]. Given this aging population, one can expect to see certain trends, such as increased numbers of TBI due to falls in the elderly, and among TBI patients, increased use of antiplatelet and anticoagulation medications [ 6 , 7 , 8 ].…”

Section: Introductionmentioning

confidence: 99%

Comparative Analysis of Antithrombotic Therapy Outcomes in Mild Traumatic Brain-Injury Patients: A Focus on Bleeding Risk and Hospital-Stay Duration

Desai,

Shiffer,

Giordano

et al. 2024

Life

View full text Add to dashboard Cite

Background: Traumatic brain injury (TBI) in the elderly is a noteworthy pathology due to the exponential increase in population age, and the effects of antiplatelet and anticoagulation on patients’ outcomes are still a matter of dispute. The aim of the present study was to evaluate the impact of various antithrombotic agents on patients with mild TBI, focusing on the risk of intracranial bleeding (ICH) and length of hospitalization (LOS). Methods: A retrospective analysis was conducted, including patients with a diagnosis of TBI admitted to the Emergency Department between 2021 and 2022. Patients were classified according to the concurrent antithrombotic therapy as aspirin (ASA), antiplatelets, direct oral anticoagulants (DOACs), and low-molecular-weight heparin (LMWH). The primary outcome was the ICH occurrence, while the secondary outcome was the LOS. The statistical analysis was performed via logistic regression models in R and STATA 13.1 software. Fisher’s exact test was used for the statistical significance. Results: 267 patients with mild TBI were included; 148 were not on antithrombotic agents, 43 were on aspirin, 33 on DOACs, 5 on LMWH, 22 on antiplatelets, and 16 on VKA. Out of the total, 9 patients experienced ICH, none of which were on DOACs, LMWH, or VKA, but 4—out of 65—were on antiplatelets, and 5—out of 148—were not on antithrombotic therapies. Patients not on antithrombotic therapy had the shortest LOS at 0.46 days, while those on VKA had the longest LOS at 1.19 days; similar trends were observed for patients on DOAC and LMWH. Conclusions: The results reveal that TBI patients on anticoagulants/antiplatelets had longer hospital stays compared with those on aspirin alone. Notably, VKA was the strongest predictor for an extended LOS. Regarding ICH, patients taking only aspirin were twice as likely to experience bleeding compared with those on anticoagulants/antiplatelets. However, to achieve statistically significant evidence, further research with a larger cohort of patients is needed.

show abstract

Word2vec Word Embedding-Based Artificial Intelligence Model in the Triage of Patients with Suspected Diagnosis of Major Ischemic Stroke: A Feasibility Study

Cited by 7 publications

References 35 publications

Natural Language Processing to Extract Meaningful Information from a Corpus of Written Knowledge in Breast Cancer: Transforming Books into Data

Natural Language Processing to Extract Meaningful Information from a Corpus of Written Knowledge in Breast Cancer: Transforming Books into Data

MIBPred: Ensemble Learning-Based Metal Ion-Binding Protein Classifier

Comparative Analysis of Antithrombotic Therapy Outcomes in Mild Traumatic Brain-Injury Patients: A Focus on Bleeding Risk and Hospital-Stay Duration

Contact Info

Product

Resources

About