In biomedical texts, abbreviations are frequently used due to their inclusion of many technical expressions of some length. Accordingly, appropriate recognition of abbreviations and their full form pairs is essential task in automatic text processing of biomedical documents. However, unlike biomedical literatures, clinical notes have many abbreviations without full form indicated in the text or without standard definition in dictionaries due to the nature of the documents. This causes difficulties in adapting traditional approaches for abbreviation disambiguation such as classification among fixed candidates or pattern-based definition extraction. Because of this reason, we consider the task as search problem and propose an approach with two steps: a) exploring possible full form candidates from various resources and b) choosing most acceptable one among retrieved candidates by ranking. To discover full form candidates and extract features of them, we exploited external academic resources such as MEDLINE and UMLS as well as clinical note corpus itself. To rank the candidates properly by consulting human criteria, we adopted RankBoost, one of learning to rank models developed from information retrieval and machine learning societies. Results show the suggested two-step approach has potential on this kind of task and propose another possible application of learning to rank models.