Legal Issue Spotting - First Phase Legal Analysis Recommendations

Legal analysis is a multistep process that performs the complex tasks of disambiguation of comported real world activities through the lens of a predefined legal code and sets of laws. These activities often may instantiate multiple separate as well as overlapping legal stipulations and criteria thus creating a potential for contextually significant interpretations of how a given code or law applies to specified events. Thus when attempting to achieve a legal resolution given ambiguous circumstances or in a competitively incentivized resolution environment, the employment of large amounts of resources is not unheard of. Despite this market incentive legal analysis has yet to be materially aided by the employment algorithmic reasoning tools, either by NLP or syntactic learning algorithms (though search and visualization tools have proved useful as research aids). We comparatively employ Naïve Bayes and SVM models in a parallel, non-mutually exclusive alignment towards the generic first phase of legal analysis: legal issue spotting. We do this by training separate supervised learning models against the digested legal opinions of the U.S. Federal Reporter, 3rd Series (F.3d) towards the U.S. Legal Code laws and sub laws to produce testable models capable of ingesting free form scenario descriptions and producing legal code research area recommendations.