RM-EMBA-BA-19

RM-EMBA-BA-19

MBA(Exe.)-BA-2019-20: Term-II

Regression Modeling

Credits	3
Program and Term	EMBA (Business Analytics), Term II
Faculty Name	Sandipan Karmakar

Course Description
Regression modeling forms the base of Predictive Analytics, which constitutes the most essential component of Business Analytics. In this era of exorbitantly soaring quantity of data and its availability, a business minded professional will always try to bring out important insights from the heap of available data which may turn out to be beneficial for important decision-making purposes. In this course, aim is to pertain required knowledge for carrying out input-output modeling to ease the process of data driven inferences beneficial for business decision-making. Starting from Simple Linear Regression to Generalized Linear Models, almost all kinds of Statistical modeling of input-output relationship will be delved into, in due course of time. Along with this the attendees will be exposed to two popular FOSSs namely R and Python to aid in computational purposes implementing different Regression Modeling approaches with the help of real-life datasets.

Leaning Outcomes

At the end of the course an attendee should
1. Be able to define a business issue as a predictive analytical problem.
2. Be able to choose from various regression modeling techniques suited in a given situation
3. Be able to understand, interpret and recommend actions based on analytical output
4. Be able use R & Python for Regression Modeling

Session Plan

Session Topics/Activities

1 Introduction to Regression Modeling –
I. Basic Tasks of Regression Models
II. Examples of Tasks of Regression Modeling

2-4 Simple Linear Regression (SLR)
I. SLR Model
II. Least Squares Estimation of the parameters
III. Hypothesis Testing on parameters
IV. Interval Estimation on parameters
V. Prediction of new observations
VI. Coefficient of Determination
VII. Considerations in Regression
VIII. Estimation by Maximum Likelihood
IX. When Regressors are random

5-7 Multiple Linear Regression (MLR)
I. MLR Model
II. Estimation of Model Parameters
III. Hypothesis Testing in MLR
IV. Confidence Interval Estimation in MLR
V. Predicting new observations
VI. Extrapolation in MLR
VII. Standardized Regression Coefficients
VIII. Multicollinearity

8-9 Model Adequacy Checking
I. Introduction
II. Residual Analysis
III. PRESS Statistic
IV. Detection and Treatment of Outliers
V. Lack of Fit of the Regression Model

10 Data Transformation and Weighting to Correct Model Inadequacies
I. Variance Stabilizing Transformations
II. Transforming to Linearize
III. Analytical Methods for Selecting a Transformation
IV. Generalized and Weighted Least Squares

11 Diagnostics for Leverage and Influence –
I. Detecting and Treatment of Influential Observations
II. Leverage
III. Measures of Influence
IV. Measure of Model Performance

11-12 Polynomial Regression
I. Model Building
II. Polynomial Models in one Variable
III. Nonparametric Regression
IV. Polynomial Models in Two or More variables
V. Orthogonal Polynomials

13 Indicator Variables
I. Concept of Indicator variables
II. Regression Approach to ANOVA

14-15 Variable Selection and Model Building
I. Stepwise, Forward, Backward and Best Subset Regression
II. Shrinkage based Variable Selection
III. Multicollinearity-Sources, Effects and Diagnostics
IV. Partial Least Squares, Quantile Regression
V. Support Vector Regression, Principal Components Regression

16 Generalized Linear Models
I. Logistic, Multinomial and Ordinal Logistic Regression
II. Poisson Regression
III. Negative Binomial Regression
IV. Cox Regression
V. Softmax Regression

17 Nonlinear Regression Models
I. Nonlinear Least Squares
II. Transformation to a Linear Model
III. Parameter Estimation
IV. Hypothesis Testing

18 Other Topics
I. Regression with Autocorrelation Errors
II. Effect of Measurement Errors in Regressors
III. Bootstrapping in Regression
IV. CART
V. Neural Networks
VI. Designed Experiments for Regression
VII. Validation of Regression Models-Overfitting, Underfitting, Cross Validation

19 Project Presentation

20 Doubt Clearing and Wrap up

Prerequisites

1. Through understanding of Basic Statistics and Linear Algebra
2. Basic Differential Calculus
Evaluation

1. 2 Quizzes of 15% + 15% weight. (Best Two out of Three) – Online
2. Mid Term of 30% weight – Offline
3. End-term of 40% weight – Offline
No Makeup Exams. Marks for missed components will be equated to the minimum of the other attended components (in percentage terms).

Reading

1. Introduction to Linear Regression Analysis by Douglas C Montgomery, Elizabeth Peck & G Vining, Wiley
Academic Integrity

Malpractice in any form will be dealt with as per manual of policies

Created By: Alora Kar on 06/26/2019 at 10:26 AM
Category: MBA(Exe.)BA-2019-20 T-II Doctype: Document

...........................

Session	Topics/Activities
1	Introduction to Regression Modeling – I. Basic Tasks of Regression Models II. Examples of Tasks of Regression Modeling
2-4	Simple Linear Regression (SLR) I. SLR Model II. Least Squares Estimation of the parameters III. Hypothesis Testing on parameters IV. Interval Estimation on parameters V. Prediction of new observations VI. Coefficient of Determination VII. Considerations in Regression VIII. Estimation by Maximum Likelihood IX. When Regressors are random
5-7	Multiple Linear Regression (MLR) I. MLR Model II. Estimation of Model Parameters III. Hypothesis Testing in MLR IV. Confidence Interval Estimation in MLR V. Predicting new observations VI. Extrapolation in MLR VII. Standardized Regression Coefficients VIII. Multicollinearity
8-9	Model Adequacy Checking I. Introduction II. Residual Analysis III. PRESS Statistic IV. Detection and Treatment of Outliers V. Lack of Fit of the Regression Model
10	Data Transformation and Weighting to Correct Model Inadequacies I. Variance Stabilizing Transformations II. Transforming to Linearize III. Analytical Methods for Selecting a Transformation IV. Generalized and Weighted Least Squares
11	Diagnostics for Leverage and Influence – I. Detecting and Treatment of Influential Observations II. Leverage III. Measures of Influence IV. Measure of Model Performance
11-12	Polynomial Regression I. Model Building II. Polynomial Models in one Variable III. Nonparametric Regression IV. Polynomial Models in Two or More variables V. Orthogonal Polynomials
13	Indicator Variables I. Concept of Indicator variables II. Regression Approach to ANOVA
14-15	Variable Selection and Model Building I. Stepwise, Forward, Backward and Best Subset Regression II. Shrinkage based Variable Selection III. Multicollinearity-Sources, Effects and Diagnostics IV. Partial Least Squares, Quantile Regression V. Support Vector Regression, Principal Components Regression
16	Generalized Linear Models I. Logistic, Multinomial and Ordinal Logistic Regression II. Poisson Regression III. Negative Binomial Regression IV. Cox Regression V. Softmax Regression
17	Nonlinear Regression Models I. Nonlinear Least Squares II. Transformation to a Linear Model III. Parameter Estimation IV. Hypothesis Testing
18	Other Topics I. Regression with Autocorrelation Errors II. Effect of Measurement Errors in Regressors III. Bootstrapping in Regression IV. CART V. Neural Networks VI. Designed Experiments for Regression VII. Validation of Regression Models-Overfitting, Underfitting, Cross Validation
19	Project Presentation
20	Doubt Clearing and Wrap up