Linear Regression
Alejandro Muñoz - A01377502
Topic
Short Definition
Application to the
workplace (Supply Chain)
Simple linear regression
Is a method to model the relationship between 2
variables. Both are independent variables. An
equation will be obtained which will show how
related this variables are.
Linear regression could be
used to explain the
relationship between demand
and seasonality to develop
better forecasting methods.
Múltiple linear regression
Is a method to model the relationship between 2 o or
more independent variables. The equation will
predict the dependent variable nsidering the value of
the other variables.
Now a model with
seasonality, material pricing
and other variables may be
used to determine the best
materials to push production
for at certain time of the year.
ANOVA
Is the analysis of variance comparing the means to
know if variables are significant. It can be obtained
from statistical software and can show the P value
for Hyptohesis testing.
Could be used to determine
optimal cost for
transportation considering
different suppliers, routes,
etc.
Qualitative variables
Dummy variables. When there’s no actual numerical
numbers such as colors, genders, etc. Boolean
values most be used to perform the regression
analysys correctly and analyze significance.
Adding customer segments or
locations to the multiple
linear regression model may
provide more data for robust
decision making.
Multicollinearity
High correlation between the independent variables.
Reviewing this is important to ensure an accurate
model
Using more variables in the
model to avoid this issue is
needed for a better model to
optimize cost within the
supply chain.