An IR-CBR approach to legal indexing and retrieval in bankruptcy law

We present and discuss the use of a combined case based reasoning and information retrieval approach to legal text indexing and retrieval in bankruptcy law. It is based on how lawyers acquire represent and retrieve information in the process of legal research and reasoning. We view the hybrid approach as a natural simplification to the intense and practically impossible to automate life-long human knowledge acquisition. We suggest and use a simple and suitable indexing and retrieval method that is based on actual statute text. We assume that a subject domain has its limited set of concepts, represented using limited sets of terms, structures, and stereotypical scenarios. We applied our model to bankruptcy case law, demonstrating its practicality and utility in areas beyond the realm of logic. Our prototype system takes as input a natural language description of a novel situation and outputs a list of similar cases.