Strengths: Clear motivation. Covariate time series could be useful for directing the time series analysis toward the analysis goals.

Points for consideration:

  1. Test for normality after detrending/deseasonalizing. The distribution of residuals is typically of more modeling concern than that of the raw data.

  2. Typos:

    • \(\epsilon\in\) should be \(\epsilon\sim\)
    • \(\psi_p\) should be \(\psi_q\).

  3. Taking logs cannot make a process stationary.

  4. AIC cannot readily be used to compare differencing - just to compare different models for the same data.

  5. Missing statement on contributions within the group, requested in the project description.