4.BDT: Machine Learning with Big Data

Code
L.24299
Amount of hours required
140
Quartile of execution
2

In this module we discuss two topics from the field of machine learning: recommendation systems and deep learning. In the first four weeks, the student pays attention to recommendation systems and metrics, and builds recommendation systems for both content-based and collaborative filtering approaches. Deep learning is covered in the second part of the module. Students have to train and test several neural networks with both fully connected layers and convolutional layers.

Competences

HBO ICT 13.2 SW/ONT/2
HBO ICT 14.3 SW/REA/3
HBO-ICT 27.3 PS/OP/OPL

Learning goals

The student understands text learning and is able to apply techniques such as Bag of Words, TF-IDF, stemming and stop words removal.
The student can implement content-based recommendation strategies for specific product data.
The student can implement both user-user and item-item collaborative filtering on a dataset with user ratings.
The student understands the ALS algorithm for building large-scale recommendation systems and can apply this using Spark ML.
The student can implement linear regression models using TensorFlow.
The student can train and test simple neural networks with fully connected layers and drop-out layers.
The student can train and test neural networks with convolutional and pooling layers.
The student is able to analyse training and test results, and van apply techniques to prevent overfitting.

Tests

Code	Name
T.50354	4.BDT: Machine Learning with Big Data
WC	Werkcollege