Premonition, which describes itself as home to the world's largest litigation database, trained a machine learning system to analyze litigation data alongside stock market data. By building regressions and quantitative models around that combined dataset, the system estimates relative stock market performance for a given industry four to twenty-four months in advance.
The underlying litigation inputs include case results and durations, analysis of currently open cases, geographic distribution of filings, and company-level litigation metrics. Premonition frames the resulting signal as distinct from traditional post-earnings analysis, positioning it as a forward-looking indicator rather than a reaction to reported earnings.
The report also touches on how this analysis can inform practical timing decisions, such as when a position might be opened and roughly how long it should be held, based on the litigation patterns the model detects.
