Develop a Holt-Winters model for forecasting temperature using R programming language. Train the model on the 1999-2000 data. Let R optimize the smoothing parameters and initial values. Use this model to generate monthly forecasts for 2001. Generate 95% prediction intervals for the 2001 forecasts. Plot the actual values (1999-2000) and the 2001 forecasts on a graph. Calculate and report forecast accuracy measures for the training data and the testing data.