DBATU · Final Year B.Tech Artificial Intelligence & Machine Learning · Non-NEP
Time Series Analysis and Forecasting
BTAIPE704A · Elective I · Semester VII · 4 credits · Continuous assessment 20 + mid-semester 20 + end-semester 60
One of the Elective I options.
Unit 1: Introduction to Time Series Analysis
7 hoursTime series and forecasting, data types and the internal structure of a series; models for analysis; autocorrelation and partial autocorrelation; the nature, uses, process and data of forecasting.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Time series and forecasting
- Components of a time series
- Autocorrelation and partial autocorrelation
- The forecasting process
Unit 2: Statistics Background for Forecasting
8 hoursTime-series plots and smoothed data; numerical description; transformations and adjustments; a general approach to modelling; evaluating and monitoring forecast performance.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Plotting and smoothing time series
- Transformations and adjustments
- Evaluating and monitoring forecasts
Unit 3: Time Series Regression Models
7 hoursLeast squares in linear regression, inference and prediction; model adequacy and variable selection; generalised and weighted least squares; regression for time-series data; first- and second-order exponential smoothing.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Least squares regression for forecasting
- Model adequacy and variable selection
- Generalised and weighted least squares
- Regression with autocorrelated errors
- Exponential smoothing
Unit 4: ARIMA Models
7 hoursARMA models, stationarity and invertibility; detecting non-stationarity with the variogram; ARIMA and seasonal ARIMA forecasting; choosing the best model and comparing impulse response functions.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- ARMA models, stationarity and invertibility
- Detecting non-stationarity
- ARIMA forecasting
- Seasonal ARIMA models
- Model selection criteria
Unit 5: Multivariate Time Series Models and Forecasting
7 hoursMultivariate stationary processes; vector ARIMA and VAR models; neural networks for forecasting; spectral analysis; Bayesian forecasting.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Vector autoregression (VAR) models
- Neural networks for forecasting
- Spectral analysis of time series
- Bayesian forecasting
Unit contents are summarised from the official DBATU syllabus . AiCanCode.org is not affiliated with Dr. Babasaheb Ambedkar Technological University (DBATU).