Lecture 5 OLS & OVB

Поделиться
HTML-код
  • Опубликовано: 7 фев 2025

Комментарии • 6

  • @arpinikoyan9076
    @arpinikoyan9076 2 года назад

    Thank you!!!

  • @ХотамжонТошматов-ч6ц

    Where can I find dataset you are using?

  • @jianleichen7750
    @jianleichen7750 2 года назад

    Question, what the relationship between OVB and correlation between feature, usually you don't put two very similar feature in linear regression, otherwise the coefficient won't make sense. (for example experience vs years since graduate. I would expect the two being more than 90% correlated, but does it suggest we should add years since graduate?)

    • @gautamsethi3751
      @gautamsethi3751 2 года назад

      You're thinking of collinearity (or multicollinearity). Collinearity and OVB are two distinct issues. For OVB, an explanatory variable X1 is correlated with the error term, and its estimated coefficient b1 will biased. If there is collinearity, then two explanatory variables X1 and X2 are correlated. If these are uncorrelated with the error term, their estimated coefficients b1 and b2 will be unbiased but their standard errors will be large.

  • @mjwork5138
    @mjwork5138 4 года назад

    Thank you very much

  • @marycrawley4721
    @marycrawley4721 4 года назад

    thank you !!