Update learning outcomes to match course content - #611

This PR closes #534.

Cross-checked each session against reference/learning_objectives.qmd and updated to reflect what is actually delivered.

Changes:

• R / statistical concepts: added Bayesian inference + MCMC and posterior interpretation outcomes (the session now contains the MCMC callout from Add A Callout for MCMC Process in "Probability Distributions and Parameter Estimation** session #383).
• Joint nowcasting with an unknown reporting delay: new subsection — the whole joint-nowcasting.qmd session had no LO entry.
• Forecasting split into Forecasting concepts (paradigm added) and Forecasting models (mechanistic↔statistical spectrum, adding structure) to match the two distinct sessions.
• Evaluating forecasts: added WIS components and PIT histograms.
• Ensemble models: added weighted ensembles (inverse-WIS, QRA).

Existing wording for delays / biases / convolution / Rt sessions left unchanged — content matches well. Methods-in-real-world / examples-of-available-tools sessions treated as supplemental and not given separate LOs.