White Noise Testing (TS E12)

The final analysis and test for time-series is White Noise. White noise is the testing of the residuals (errors) to see if any structures or features are missing from the time-series model. Three features of the residuals must exist to have White Noise. 1) Zero Mean 2) Constant Variance 3) Serial Correlation 1 and 2 can be tested with a zero mean ADF (stationarity) test. 3 can be testing with the Breusch-Pagan test for models with MA terms and a DW test for models without a MA term. An ACF of the residuals is also a good way to test and visually review the serial correlation. The ACF will help determine missing model components. Remember that the residual must be White Noise AND the dependent and independent inputs need to be stable (stationary). ....BOOKS.... Statistical Analysis of Financial Data in R: https://amzn.to/2PrRpGN Statistical Analysis of Financial Data in S-Plus: https://amzn.to/2ToFStk Quant t-shirts, mugs, and hoodies: https://teespring.com/stores/fancy-quant Connect with me:   / dimitri-bianco     / dimitribianco   ☕ Show Your Support and Buy Me a Coffee ☕ https://ko-fi.com/fancyquant