Representation Learning for Customers

2018-2019 Spring
Faculty Department of Project Supervisor: 
Faculty of Engineering and Natural Sciences
Number of Students: 

Representation learning has emerged as an alternative approach to feature extraction in machine learning. In representation learning,
features are extracted from unlabeled data by training a deep neural network on a secondary, supervised learning task. The most popular
example is the word2vec approach, which allows capturing the meaning of the words in multi-dimensional vectors. In this project, the
students will apply a similar strategy to represent customers of a business. By analyzing the purchase history, the customers will be
embedded in a vector space. These vectors will be useful in many other subsequence analyses such as churn prediction or loyal customer prediction tasks.

Related Areas of Project: 
Computer Science and Engineering
Visual Arts and Visual Communications Design