To predict the clickthrough rates using models for the data provided by CriteoLabs as a part of the Kaggle challenge

Abstract

Advertisements displayed on the website are a major source of revenue for website/blog owners today. It is one of the central applications of Machine Learning Techniques. Clickthrough Rate (CTR) represents the ratio of the number of times the ad is clicked upon divided by the total of times the ad is displayed on the website. CTR = 𝑛𝑜. 𝑜𝑓 𝑡𝑖𝑚𝑒𝑠 𝑎𝑑 𝑖𝑠 𝑐𝑙𝑖𝑐𝑘𝑒𝑑 𝑛𝑜. 𝑜𝑓 𝑖𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛𝑠 X 100 (%) This is an important metric to be tracked in Pay-Per-Click (PPC) Models where the revenue generated depends on the clicks which the advertisement receives. CTR shows how well an ad is appealing to the website visitor and whether they need to change some keywords. A high number of click through rates shows that user finds the ad relevant and marketing campaign was effective.