You are on page 1of 3

PREDICTIVE

ANALYSIS IN
RETAIL
By- Prashant Paliwal

AIM- To predict the retail in upcoming


months depending upon existing monthly
retail data using machine learning.
APPROACH

DATA
We will have the retail data of HONDA MOTORS sorted Day or Month-Wise along with dealer code, model
no., model name of respective product. As shown in the table below:

Model Model Model


Month Dealer Code Colour Retail Count
Code Name Variant

STEP 1:
Data Preprocessing
This stage involves feeding of given data into our program and performing operations like dimensionality
reduction, scaling, finding of any missing values.

After this Class Labels and Features and determined which in this case are Model Code and Retail Count
respectively.

In this stage the given data is divided into train and test sets using sci-kit learn library in ml.

STEP 2:
Learning Algorithm
Here, we will do Model Selection based on our data and compare it with different models in order to get
best possible prediction as output.

Since this is a case of supervised learning so various regression models can be chosen to perform the given
predictive analysis.

Here, we will compare Linear, Logistic and multiple regression and use the most suitable one to predict the
output.

Page 1
APPROACH

STEP 3:
Prediction
In this stage since we have trained our learning algorithm with existing data will give new data or in this
case next month as input in order to predict the output that is the retail count of input month depending
upon our algorithm.

Here, multiple epochs will be done and compared in order to give the accuracy of our predictor model.

Hence, helping the existing dealers to compare their actual retail with predicted retail.

Page 2

You might also like