In recent years, deep learning has significantly transformed ventilator pressure forecasting , which is crucial for providing patients with safe, personalized respiratory therapy. This work introduces a novel hybrid model that combines Deep Neural Networks (DNN) with Long Short-Term Memory (LSTM) networks. To enhance forecast accuracy, we incorporated a special forget gate reset mechanism to this model. Our approach, which involved meticulous data collection and analysis, produced a robust model architecture that effectively captures the unique characteristics of each breath cycle. We examined various data-splitting techniques, particularly comparing Timeseries Split and K-fold cross-validation to determine the most effective one. According to our findings, Timeseries Split performs better at preserving the sequential order of the ventilator data. Notably, compared to traditional methods, our hybrid model achieved an astounding 84% reduction in Mean Absolute Error (MAE) and a 97 % reduction in Mean Squared Error (MSE). These results highlight the potential of our approach to greatly improve ventilator control by making precise, data-driven predictions.
Competing Interest StatementThe authors have declared no competing interest.
Funding StatementThe author(s) received no specific funding for this work.
Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Not Applicable
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
The research described in this manuscript involves the use of publicly available data from Kaggle and does not involve human participants, human specimens, vertebrate animals, or field research. Therefore, IRB or ethics committee approval was not required for this study.
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Not Applicable
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Not Applicable
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Not Applicable
Data AvailabilityThe data underlying the findings described in this manuscript are fully available without restriction. The dataset used in this study is publicly accessible on Kaggle at the following URL: [https://www.kaggle.com/competitions/ventilator-pressure-prediction/data]. The ventilator data, produced using a modified open-source ventilator connected to an artificial bellows test lung via a respiratory circuit, contains control inputs and the state variable (airway pressure) to predict. This dataset was utilized to train and evaluate the proposed Hybrid LSTM-DNN model for ventilator pressure prediction.
https://www.kaggle.com/competitions/ventilator-pressure-prediction/data
Comments (0)