Clear-air turbulence (CAT) is a significant safety issue for aviation at upper levels (>=20000ft mean sea level) in the atmosphere. CAT forecasting products, (e.g., Graphical Turbulence Guidance (GTG), one of the most reliable systems (Sharman et al., 2006)) use a global optimization technique to produce a forecast from many different indicators, or diagnostics, of turbulence. Although it is intentively obvious that many diagnostics vary in their predictive skill depending on geographic region or altitude band, GTG is unable to exploit these regional dependencies due to an insufficient number of timely Pilot Reports (PIREPs) for real-time evaluation of individual diagnostic skills. The In-situ Reporting System has provided more plentiful and objective observation data in recent years, enabling a regional approach and the opportunity to use machine learning algorithms such as Support Vector Machines, Random Forests and logistic regression to improve turbulence forecast accuracy. This paper compares the accuracy of these methods, prototypes regionalization strategies using machine learning which reflect the climatology of CAT, and explores approaches to combining regions to form a coherent product for future operational CAT forecasting.