Introducing Regression Models in SAP Analytics Cloud Smart Predict

Objectives

After completing this lesson, you will be able to:

Explain regression analysis in Smart Predict

Use Cases for Regression Models

Use Augmented Analytics for Payment Forecasting

In the case study, we walk you through a scenario for using regression models in payment forecasting.

What sorts of topics can we investigate with a regression model?

Measuring how an increase in the costs of a product impacts a company’s profits.

Understanding how sensitive a company’s sales are to changes in advertising spend, promotions, or pricing.

Analyzing how a change in interest rates affects a stock price.

Predicting accident and loss claim values for a car insurance company, based on factors such as the attributes of the car, driver information, and demographics.

Forecasting the future consumption of electricity, based on historical demand, weather forecasts, and pricing.

Regression Analysis in Smart Predict

What is regression analysis?

Regression analysis is a collective name for techniques used for the modeling and analysis of numerical data consisting of values of a target variable and of one or more influencer variables.

The parameters of the regression are estimated and give a "best fit" of the data.

The target variable in the regression equation is modeled as a function of the influencer variables, a constants term, and an error term. The target is a continuous variable.

Regression Lines

The formula for a simple regression line is represented as an equation: y = a + bx.

Where:

y is the target.

a is the intercept (the level of y where x is 0).

b is the slope of the line.

x is the influencer variable.

Multiple Linear Regression

Multiple linear regression is used to explain the relationship between one continuous target variable and two or more influencer variables.

The influencer variables can be continuous or categorical.

Multiple linear regression analysis is the task of fitting a single line through a scatter plot, with multiple dimensions of data points.

Regression is most often used to:

Identify the strength of the effect that the influencer variables have on a target variable.

Forecast effects or impacts of changes - to understand how much the target variable changes when you change the influencer variables. For example, a multiple linear regression can explain how much sales volumes are expected to increase (or decrease) for every one-point increase (or decrease) in workforce hours.

Predict trends and future values. The multiple linear regression analysis can be used to get point estimates. An example question is: What will the price of gold be six months from now?