Fuzzy Linguistic Concept Rule Extraction Algorithm for Ventilator-Associated Complications in Patients with Acute Respiratory Failure
Ning Kang, Zhichen Zhu, Ziyi Jin, Yixiang Chen · 2024
Ventilator-Associated Condition (VAC) describes lung damage that arises in patients undergoing mechanical ventilation. This includes newly developed or aggravated pulmonary infiltrates, bronchitis, pneumonia, and other similar conditions. This paper primarily investigates a fuzzy reasoning method for ventilator-associated complications in patients with acute respiratory failure. It presents a rule extraction algorithm based on fuzzy linguistic concepts, aiming to perform fuzzy reasoning on whether patients are likely to develop ventilator-associated conditions (VAC) based on linguistic concepts. Firstly, the importance of features is evaluated using the Out-Of-Bag (OOB) method, and the top five features with the highest importance ranking are selected as the core attributes for this study: peep, spo2, resp-rate-total, platelet, and vt-PW. Next, appropriate fuzzy membership functions are established by fitting the distribution frequencies of the core attributes. This forms the foundation of the fuzzy linguistic decision context. Then, by considering the refined relationships between fuzzy linguistic concepts based on conditional and decision attributes under different levels of confidence, personalized linguistic rules are extracted to capture individualized requirements. Finally, the effectiveness of the proposed method is demonstrated through an example. This research provides a scientific basis for uncertain reasoning and decision-making with linguistic value information.