https://github.com/chitranjan806/advertisement_success_prediction

A Binary Classification problem to predict whether the revenue generated will cover costs to produce and air the ad(Whether there will be a net gain from an ad or not).
https://github.com/chitranjan806/advertisement_success_prediction

Last synced: 4 days ago
JSON representation

A Binary Classification problem to predict whether the revenue generated will cover costs to produce and air the ad(Whether there will be a net gain from an ad or not).

• Host: GitHub
• URL: https://github.com/chitranjan806/advertisement_success_prediction
• Owner: Chitranjan806
• License: mit
• Created: 2020-12-15T05:42:52.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2020-12-15T12:10:24.000Z (almost 4 years ago)
• Last Synced: 2023-09-05T17:47:38.442Z (about 1 year ago)
• Language: Jupyter Notebook
• Size: 198 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Advertisement_Success_Prediction

A Binary Classification problem to predict whether the revenue generated will cover costs to produce and air the ad(Whether there will be a net gain from an ad or not).

This is an Assessment Task, and thus, the Dataset(s) is not shared in the repository.

## Overview

The holiday season is just around the corner—Christmas trees have been decorated, lights and wreaths hung, streets all decked up, Santa costumes rented out, and holiday cards in the mailbox.

Because of holiday cheer, retail brands, big and small, want to earn considerable profits, and therefore, are investing significantly in advertising. These brands have approached an advertising agency to plan and execute ad campaigns that will help them increase the footfall in their stores.

You have been hired by this advertising company to assess the revenue that can be generated by a proposed ad. Based on the demographic information provided, you need to predict whether the revenue generated will cover costs to produce and air the ad(Whether there will be a net gain from an ad or not)

This will help guide decision-making for the firm, as they will want to pursue ads that are likely to generate a net gain for their clients— thereby boosting the advertising firm’s reputation.

## Evaluation Metric

Submissions are evaluated using [F1_Score(Binary)](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html)

## Results

### (Based on F1 Scores)

|Algorithm Used          |On Train data |On Validation data|

|:-----------------------|------------:|----------------:|

|***Logistic Regression***     |0.49516       |0.49668           |

|***Decision Tree Classifier***|0.69203       |0.51972           |

|***Cat Boost Classifier***    |0.61420       |0.53378           |

|***XGBoost Classifier***      |0.54720       |0.54442           |

As observed from the above table of F1 Scores, the ***best results were obtained using XGBoost Classifier*** algorithm. Thus, the predictions saved as ***'submission.csv'*** is performed using XGBoostClassifier.