Scribd
Upload
21 views3 pages

ITXXX Applied Forecasting Methods Winter - Pritam Anand

The course 'Applied Forecasting Methods' (SC475) is a technical elective worth 4 credits, focusing on forecasting models from a machine learning perspective. It covers a range of topics including linear regression, quantile regression, time series models, and deep learning techniques, with practical applications in finance, energy, and sustainability. Students will gain hands-on experience through lab classes and develop skills in modern forecasting methods used in research and industry.

Uploaded by

202418003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
21 views3 pages

ITXXX Applied Forecasting Methods Winter - Pritam Anand

The course 'Applied Forecasting Methods' (SC475) is a technical elective worth 4 credits, focusing on forecasting models from a machine learning perspective. It covers a range of topics including linear regression, quantile regression, time series models, and deep learning techniques, with practical applications in finance, energy, and sustainability. Students will gain hands-on experience through lab classes and develop skills in modern forecasting methods used in research and industry.

Uploaded by

202418003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

1.

Course Title Applied Forecasting Methods


Lecture hours per week: 3
Tutorial hours per week: 0
2. Credit Structure
Practical hours per week: 2
Total Credits: 4
3. Course Code SC475
B.Tech. (ICT) (VI and VIII Semester) / M.Tech. (ML) (II Semester) Winter
4. Program/Semester
Semester
5. Category Technical Elective
6. Prerequisite
Machine Learning, Basic Programming.
courses
7. Foundation for UG (VI Semester) / Industry Applications that involves forecasting tasks
This course aims to develop a systematic understanding of the forecasting
models from machine learning perspective. The course offers a smooth journey
from basic Least Square Regression model developed around year 1805 to
8. Abstract Content
modern complex forecasting methods after 2020. The content of the course
covers the variants of point and probabilistic forecasting methods with detailed
statistical and mathematical analysis for panel and time-series data.
(i) Hyndman, R.J., & Athanasopoulos, G. (2018) Forecasting:
principles and practice, 2nd edition, OTexts: Melbourne,
Australia. OTexts.com/fpp2 .

(ii) Zhang A ,Lipton ZC, Li M and Smola AJ , (2023). Dive into


deep learning. Cambridge University Press.

Text Book and (iii) Koenker, R., Chernozhukov, V., He, X., & Peng, L. (Eds.).
Materials (2017). Handbook of Quantile Regression (1st ed.). Chapman
and Hall/CRC. https://doi.org/10.1201/9781315120256 .

(iv) Takeuchi, I., Le, Q., Sears, T., & Smola, A. (2006).
Nonparametric quantile estimation, 2006, Journal of Machine
Learning and Research , vol. 7, pp. 1231–1264.

Page 1 of 2
Detailed Course Contents
Topic Name Content (2 -3 lines per 4 – 6 lectures) No. of lectures
Application of forecasting in different verticals like energy,
Introduction to
finance, environment and health. Motivation, Definition, 2
Forecasting Methods
objectives and approaches in Forecasting,

Linear regression models, Least Square Methods, Non-linear


Regression from ML
regression models, Kernel Regression, Gradient Descent 6
perspective
Methods, Evaluation Metrics.

Quantile Regression model, Pinball loss function, Quantile


Quantile Regression & Regression with Kernel Machine, Quantile Regression
Uncertainty Neural Network, Prediction Interval Estimation, Evaluation 12
Quantification Metrics, Uncertainty Modeling with Kernel Machine and
Neural Network.

Statistical Time-series Introduction to Time Series Forecasting , AR, ARMA and


8
Models ARIMA models.

Time Series Forecasting with Recurrent Neural Network


Time Series
(RNN), Gated Recurrent Unit (GRU), Long Short Term
Forecasting with Deep 8
Memory (LSTM) , Deep Auto Regressive (Deep AR)
Learning
model, Time GPT model.
Probabilistic Probabilistic Forecasting in Neural Network, RNN, LSTM
4
Forecasting and Deep AR models.
- Systematically develop a deep understanding of the main principles of
forecasting.
- Develop the machine learning perspective for forecasting and uncertainty
modeling.
Outcomes and - Develop the understanding of statistical and modern time series models for
Objectives point and probabilistic forecasting tasks.
- Develop the implementation and practical skills for modern and state-of art
forecasting methods used in research and industry after COVID 19.
- Hands on experience for different forecasting tasks in Finance, Energy, and
Sustainability verticals.
Lab – The Lab classes would be dedicated for systematic implementation of
algorithms for forecasting tasks in Finance, Renewable Energy, Electricity Load,
Environment and sustainability.
Use cases for Lab class: -
Comments  Benchmark Datasets.
 Electrical Load Datasets.
 Renewable Energy (Wind and Wave) Datasets
 Financial and Stock Prices Datasets.
 Environment and Pollution Datasets.
POs-COs Matrix:

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12

x x x x x

Page 2 of 2

You might also like