STATS 499 Lecture Notes - Lecture 9: List Of Fables Characters, Descriptive Statistics, Google Scholar
Document Summary
Maybe the first few years this predictor was accurate, and that was the reason this predictor was proposed. After you find a predictor, you should not include the initial data. Exclude the initial period and look into the testing period only. Your accuracy would be 70% if you always predicted up. Also, you could find other published papers to cite: misc. If you talk about forecasting power, describe what you mean by forecasting power. Spurious correlation make that quantifiable, don"t just state what it is and the potential for it to occur. Example) a non-smokers compared to a smokers" risk of lung cancer. Risk lung cancer in a smoker is 7 times as high as a non-smoker: almost all studies talk about relative risks, but you can always claim correlation not causation. Writing a research paper: for assignment #4, start looking at the data! The data is messy and this assignment cannot be done in a week!
