AI Tech Trainee module 3

Which of these is not necessarily true of a time series?

• Time series data are taken in equally spaced points in time

• Time series data can be ordered by time

• Time series data are a special type of financial data

• Time series data are quantitative

Which of the following is not a result of time series decomposition?

• Trend

• Cyclical component

• Seasonality

• Residuals

In time series, 'taking a lag of a time series' means looking at how well are we able to predict past data.

• True

• False

What can we learn by calculating autocorrelations?

• How closely related are two different time series

• How much do the residuals of different periods vary around the mean

• How closely related is a time series with some lag of itself

• How much more seasonality there is in the time series relative to how much trend there is

Which one of these is not a condition for a time series to be stationary?

• Constant mean

• Constant variance (homoskedasticity)

• No seasonality

• Small residuals

How do we call the group of very simple models:

• Easy models

• Simpletons

• Naïve models

• Average models

Time series model components can be usually decomposed using an additive or multiplicative approach.

• True

• False

What are the possible values of the numeric parameters (alpha, beta, gamma) in Exponential smoothing models:

• from 0 till 100

• from 0.1 till 0.9

• from 0 till 1

• from 0.1 till 10

What does the RMSE stand for?

• Root Median Sum Error

• Root Mean Sum Error

• Root Median Squared Error

• Root Mean Squared Error

Cross-validation is a technique used for identifying the best fitting model for a given time series.

• True

• False