How Machine Learning Can Predict Ads Delivery

Predicting advertising campaign delivery using Machine Learning (Random Forests)

Presented By Alessio Antonioli
Alessio Antonioli
Alessio Antonioli
Data Scientists at Shopfully/Doveconviene

Alessio is the leading Data Scientist at Shopfully/Doveconviene with a 10+ years’ experience in business problem solving based on data inference and Advanced Analytics. He has an MBA in addition to a double degree in engineering.

He uses R to create algorithms that provide ready to use solutions to business issues. At work, he is an evangelist for decisions making based on robust methodologies of applied Data Science.

He truly believes that best Data Science solutions come from exhaustive business comprehension along with Statistics and Computer Science mastering.

Presentation Description

A mobile advertising platform is a marketplace that allows promotional advertisers to meet qualified shoppers. A modern competitive platform has to guarantee customers the delivery of an established number of content views within a specific time span. To fulfill the commercial commitment it is important being able to predict the risk of missing the purchased target so to leverage some operational adjustments far in advance.

In such business cases, the variable selection process has a crucial role in building a robust machine learning model. It has to pass a ‘business validation process’ ensuring that all important variables are included.

The session will cover a full Data Science framework, from data exploration to output visualization, including the validation of the regression model.

The study will make use of a non-linear machine learning model for regression and an unsupervised classification for variable simplification.

Presentation Curriculum

Predicting advertising campaign delivery using Machine Learning (Random Forests)
45:22
Hide Content