Today in class, we learned about time series. We learned that data can depend on time. We looked at how to look at the residuals and make the model a quadratic model with time as the predictor. In some cases, you can go until the 8th degree or higher to find the best model. It works best to use the command : myMod8 <- lm(birth ~ poly(year,8) ) when making models with high powers. If you use a high power in your model, it is important to use all lower powers in your model as well. We also learned about the dwtest. This is a test that gives us a p-value in which we can use to determine whether there is autocorrelation or not. If there is auto correlation, we need to determine if there is negative auto orrelation or positive autocorrelation. Right now, all we know how to do is determine if there is autocorrelation or not. Positive Autocorrelation: Positive residual is lively to be followed by a positive residual. Negative residual is likely to be followed by a negative residual. Negative Autocorrelation: Positive residual is lively to be followed by a negative residual. Negative residual is likely to be followed by a positive residual.