Core Concepts
ICE-SEARCH integrates language models with evolutionary algorithms for effective feature selection in medical predictive analytics.
Abstract
ICE-SEARCH introduces a novel approach that combines language models with evolutionary algorithms for feature selection in medical predictive analytics. The study demonstrates the effectiveness of ICE-SEARCH in tasks such as stroke, cardiovascular disease, and diabetes prediction, outperforming traditional methods and showcasing robustness and scalability. The algorithm's adaptability with different language models and its ability to identify high-quality feature sets from large datasets are highlighted.
Stats
ICE-SEARCH는 의학 예측 분석에서 효과적인 특징 선택을 위해 언어 모델과 진화 알고리즘을 통합합니다.
ICE-SEARCH는 전통적인 방법을 능가하며 스트로크, 심혈관 질환 및 당뇨병 예측과 같은 작업에서 효과적임을 입증합니다.
ICE-SEARCH는 다양한 언어 모델과의 적응성 및 대규모 데이터셋에서 고품질 특징 집합을 식별하는 능력을 강조합니다.