We study the problem of sequentially testing individuals for a binary disease outcome whose true risk is governed by an unknown logistic model. At each round, a patient arrives with feature vector $x_t$, and the decision maker may either pay to administer a (noiseless) diagnostic test--revealing the true label--or skip testing and predict the patient's disease status based on their feature vector and prior history. Our goal is to minimize the total number of costly tests required while guaranteeing that the fraction of misclassifications does not exceed a prespecified error tolerance $\alpha$, with probability at least $1-\delta$. To address this, we develop a novel algorithm that interleaves label-collection and distribution estimation to estimate both $\theta^{*}$ and the context distribution $P$, and computes a conservative, data-driven threshold $\tau_t$ on the logistic score $|x_t^\top\theta|$ to decide when testing is necessary. We prove that, with probability at least $1-\delta$, our procedure does not exceed the target misclassification rate, and requires only $O(\sqrt{T})$ excess tests compared to the oracle baseline that knows both $\theta^{*}$ and the patient feature distribution $P$. This establishes the first no-regret guarantees for error-constrained logistic testing, with direct applications to cost-sensitive medical screening. Simulations corroborate our theoretical guarantees, showing that in practice our procedure efficiently estimates $\theta^{*}$ while retaining safety guarantees, and does not require too many excess tests.

翻译：暂无翻译