Machine Learning - Logistic Regression

Facebook--PredictCustumer-Click

Background

Running a targetted marketing ads on facebook. The company wants to anaylze customer behaviour by predicting which customer clicks on the advertisement. Customer data is as follows:

Inputs:

• Name
• e-mail
• Country
• Time on Facebook
• Estimated Salary (derived from other parameters)

Goals

• Split the data in Train and Test
• Train and Test the model in the data set
• Visualize
• Predict Click on the ad

How to run

Open Google Colab https://colab.research.google.com/

• File
• Upload Notebook
• Run the Cells

Proccess

Import the data set and visualize the data

• With Scatter plot

• With Box plot

• With Histogram

Transforming the data and Executing a training Test

With Confusion Matrix, checking on the accuracy

ploting the boundary using the trained classifier

• Run the classifier to predict the outcome on all pixels with resolution of 0.01
• Colouring the pixels with 0 or 1
• If classified as 0 it will be magenta, and if it is classified as 1 it will be shown in blue

Ploting all the actual training points

Visualising the Training set results for Tran and Test