The cpm-prediction-with-miq's intro from shiva-kumarj

cpm-prediction-with-miq's Introduction

# Capstone project with MiQ to predict the optimum CPM for Real-Time-Bidding

This is the final project in the Upgrad's Machine Learning Course.

Problem Statement: Predict the optimum CPM in RTB.

RTB is the process of selling ad impressions to advertisers in real time when the ad server receives an ad query for every customer visit.
"impressions" is one ad shown to one user on one website. Ad agencies participate in an automated bidding system for every impression and the ad by the highest bidder is the winner.
Since everyone is working with a limited budget, bidding higher the highest is not always optimal, which would lead to quick exhaustion of the budget before the target for the number of impressinos is met. 
The intelligent strategy is to gather insights of the customer behaviour from multiple sources and then use that to predict what the optimum CPM is for that impression. bidding any higher than that predicted amount will lead to diminishing returns.

The project here is to build a ML learning model to learn from the customer's historical data and then use this model to predict the CPM for new customers, the goal is to get as close to the optimum CPM as possible while remaining within the budget limit.

Recommend Projects