Data Analysis Flashcards
ggplot()
Scatterplot that shows correlation
GGally: ggpairs ()
Scatterplot matrix shows the correlation between various time series (more than 2)
Positive correlation in lag plots
Means strong seasonality
What does correlation measure?
It measures the extent of linear relationship between two variables
What does autocorrelation measure?
It measures the linear relationship between lagged values of a time series
What is a “lag”?
A fixed amount of passing time
How data changes depending on the autocorrelations for the lags?
- When data have a trend, the autocorrelations for small legs tend to be larger and positive because obdervations nearby in time are also nearby in value. Thus, the ACF of a trended time series tends to have positive values that slowly decrease as the lags increase
- When data are seasonal, the autocorrelations will be latger for the seasonal lags than for other lags
- When data are both trended and seasonal you see combination of those effects.
White noise
For white noise series we expect autocorrelation to be close to zero, 95% of spikes in the correlogram (ACF) to lie withing blue lines (+-2^T)
3 types of time series pattern?
Seasonality, trend and cycles
3 components that are comprising time series? Formula
A trend-cycle component (or trend), a seasonal component, a remainder component (containing anythjng else in the series)
yt = St +Tt + Rt
y-data
S-seasonal
Box Cox transformation purpose
It transforms the data so that it closely resembles a normal distribution. (To make errors normally distributed to construct confidence intervals and conduct hypothesis tests).
Additionally transforming our variables csn improve the predictive power of our models because transformations can cut away white noise
What is guerrero feature?
It can be used to choose the best lambda value for u in box cox transformstion
