Evaluation of Linear Imputation Based Pediatric Appendicitis Detection System Using Machine Learning Algorithm

Appendicitis is a condition that can be crucial, and lack of proper treatment among the children may cause unwanted death. The primary objective of this study is to develop a model specifically trained for diagnosing pediatric appendicitis. Detailed study has been performed by following specific steps and techniques involved in data preprocessing and model development phase. The experiment found the challenges associated with diagnosing appendicitis in children, including the presence of nonspecific symptoms, missing symptoms, and variations in clinical presentations. The researchers acknowledged the importance of addressing these challenges associated with nonspecific symptoms, missing symptoms, and variations in clinical presentations, the developed model demonstrates the potential to improve diagnostic accuracy in children. In learning that the experiment delved into the model development process, encompassing the selection of suitable machine-learning algorithms or statistical techniques like interpolation for handling missing values. The rationale behind these choices is explained, along with insights into how the model was trained and evaluated using the available dataset. The linear interpolation algorithm has been found to achieve optimal results with logistic model accuracy of 87% for the prediction of appendicitis, 95% for the prediction of appendicitis treatment, and 91% for the prediction of complications of appendicitis. However, an estimated prediction model based on ML has been built to predict pediatric appendicitis.