You are an ML engineer at a manufacturing company that specializes in producing high-value machinery. The company is facing significant expenses due to unexpected machine failures. To mitigate this, you are tasked with creating a classification model for a predictive maintenance use case. The goal is to predict whether a crucial machine will fail within the next three days, allowing the repair crew ample time to address the issue before a costly breakdown occurs. Regular maintenance of the machine is relatively inexpensive, but an actual failure leads to substantial costs. You have trained several binary classifiers to predict machine failure, where a prediction of 1 indicates an impending failure. Currently, you are evaluating each model using an evaluation dataset. Your objective is to choose a model that prioritizes early detection of failures while ensuring that more than 50% of maintenance jobs triggered by the model correspond to actual imminent machine failures. Given this scenario, which model should you choose?