lecture

3 Novembro 2025, 09:30 • Adriana Cornea-madeira

We discussed about model selection based on information criteria and regularization methods such as LASSO (using the information criteria to choose the tuning parameters). We also discussed about inference problems using the post LASSO OLS results and proposed the post double selection (debiased machine-learning) for doing correct inference on the parameter of interest. Moreover, we discussed about how to do iterative forecasts and direct forecasts. We then discussed about local projections and impulse response and inference for impulse responses. We then, discussed about tests for testing equal forecast accuracy (Diebold and Mariano, Giacomini and Rossi, West, Clark and West).  

---

lecture

27 Outubro 2025, 09:30 • Adriana Cornea-madeira

We discussed about unit root testing, diagnostic residuals checks, and HAC.

---

lecture

20 Outubro 2025, 09:30 • Adriana Cornea-madeira

We had the presentations of the second part of the project. We discussed about the impact of non-stationarity on testing and forecasting, and of the importance of unit root testing.

---

lecture

13 Outubro 2025, 09:30 • Adriana Cornea-madeira

We discussed about conditions for stationarity and invertibility, the ACF/PACF of stationary and nonstationary ARMA processes, and simulated some ARMA processes in R looking at the ACF/PACF patterns.

---

lecture

6 Outubro 2025, 09:30 • Adriana Cornea-madeira

At the beginning of the lecture we had the presentation of the first part of the project by each group representative. We then moved on introducing the ARMA models.

---