Chi-Squared Distribution

  • always positive
  • typically right-skewed
  • one param: $\text{df}$

9363c9506c8f4953973be84906eb1e6a.png


  • when df becomes larger, the center moves, i.e. the mean value grows as well
  • mean = df
  • variability increases as df increases
  • shape becomes less skewed, more symmetric
  • for Chi-Squared tests typically need only upper-tail values


Visualization

Sources

Machine Learning Bookcamp: Learn machine learning by doing projects. Get 40% off with code "grigorevpc".

Share your opinion