This paper offers a multi-stage proof model for early detection of pulmonary effusions. The assessment approach evaluates character elements related to pulmonary effusions, patient traits, scientific manifestations, and imaging capabilities. A supervised learning-based totally gadget mastering version is used to acquire, examine, and perceive patterns from the multi-stage evidence so that it will stumble on pulmonary effusions from the patient at an advance level. The evaluation of the multi-level evidence model is carried out via the usage of a massive set of scientific statistics and the usage of metrics consisting of sensitivity, precision, and f1-score. The evaluation results show that the model can come across early-stage pulmonary effusions with excessive accuracy and with an excessive diploma of consistency. The assessment additionally suggests that the multi-level proof model is higher than unmarried-degree function models in accuracy and with comparable consistency.