Let us help you find the training program you are looking for.

If you can't find what you are looking for, contact us, we'll help you find it. We have over 800 training programs to choose from.

Practical Recommender Systems

  • Course Code: Data Science - Practical Recommender Systems
  • Course Dates: Contact us to schedule.
  • Course Category: Big Data & Data Science Duration: 4 Days Audience: This course is geared for Python experienced developers, analysts or others who are intending be able to read code in a programming language such as Python or Java, you should understand an SQL query, and you should have a basic understanding of higher math and statistics.

Course Snapshot 

  • Duration: 4 days 
  • Skill-level: Foundation- Practical Recommender Systems for Intermediate skilled team members. This is not a basic class. 
  • Targeted Audience: This course is geared for Python experienced developers, analysts or others who are intending be able to read code in a programming language such as Python or Java, you should understand an SQL query, and you should have a basic understanding of higher math and statistics.  
  • Hands-on Learning: This course is approximately 50% hands-on lab to 50% lecture ratio, combining engaging lecture, demos, group activities and discussions with machine-based student labs and exercises. Student machines are required. 
  • Delivery Format: Delivery Format: This course is available for onsite private classroom presentation, or remote instructor led delivery, or CBT/WBT (by request). 
  • Customizable: This course may be tailored to target your specific training skills objectives, tools of choice and learning goals. 

Are you envious when Amazon recommends its products or when Netflix is spot-on with a recommendation for a user? Then here’s your chance to learn how to add these skills to your repertoire. Reading this course will give you an understanding of what recommender systems are and how to apply them in practice. To make a recommender work, many things need to perform in concert. You need to understand how to collect data from your users and how to interpret it, and you need a toolbox of different recommender algorithms so you can choose the best one for your particular scenario. Most importantly, you need to understand how to evaluate whether your recommender system is doing its job well. All this and more is hidden within this course. 

 Working in a hands-on learning environment, led by our Recommendation Systems expert instructor, students will learn about and explore: 

  • you should be able to read code in a programming language such as Python or Java  
  • you should understand an SQL query, and you should have a basic understanding of higher math and statistics.  
  • Figures and code listings that explain concepts can get you only so far. 

Topics Covered: This is a high-level list of topics covered in this course. Please see the detailed Agenda below 

  • How to understand users and their behavior, and covers ways to collect data from users. 
  • Introduces web analytics and shows how you can implement a dashboard where you can keep track of your recommenders. 
  • How behavioral data can be transformed into ratings. 
  • Outlines the problem of new users and products and gives simple solutions. 
  • Discusses formulas for calculating similarity between users or content items such as movies. 
  • presents a way to mix types of recommenders. 
  • Introduces ranking algorithms and methods for learning to rank recommendations. 
  • looks at non-personalized recommendations. 

Audience & Pre-Requisites 

This course is geared for attendees with Recommendation Systems with Python skills and infrastructure needed to get them up and running. Managers will find this course useful to get an overview of what a recommender system is and how it can be used in practice. 

Pre-Requisites:  Students should have  

  • Basic to Intermediate IT Skills. Attendees without a programming background like Python may view labs as follow along exercises or team with others to complete them. 
  • Good foundational mathematics or logic skills 
  • Basic Linux skills, including familiarity with command-line options such as ls, cd, cp, and su 

Course Agenda / Topics 

1. What is a recommender? 

  • Real-life recommendations 
  • Taxonomy of recommender systems 
  • Machine learning and the Netflix Prize 
  • The MovieGEEKs website 
  • Building a recommender system 

2. User behavior and how to collect it 

  • How (I think) Netflix gathers evidence while you browse 
  • Finding useful user behavior 
  • Identifying users 
  • Getting visitor data from other sources 
  • The collector 
  • What users in the system are and how to model them 

3. Monitoring the system 

  • Why adding a dashboard is a good idea 
  • Doing the analytics 
  • Personas 
  • MovieGEEKs dashboard 

4. Ratings and how to calculate them 

  • User-item preferences 
  • Explicit or implicit ratings 
  • Revisiting explicit ratings 
  • What are implicit ratings? 
  • Calculating implicit ratings 
  • How to implement implicit ratings 
  • Less frequent items provide more value 

5. Non-personalized recommendations 

  • What’s a non-personalized recommendation? 
  • How to make recommendations when you have no data 
  • Implementing the chart and the groundwork for the recommender system component 
  • Seeded recommendations 

6. The user (and content) who came in from the cold 

  • What’s a cold start? 
  • Keeping track of visitors 
  • Addressing cold-start problems with algorithms 
  • Those who doesn’t ask, won’t know 
  • Using association rules to start recommending things fast 

7. Finding similarities among users and among content 

  • Why similarity? 
  • Essential similarity functions 
  • k-means clustering 
  • Implementing similarities 

8. Collaborative filtering in the neighborhood 

  • Collaborative filtering: A history lesson 
  • Calculating recommendations 
  • Calculating similarities 
  • Amazon’s algorithm to precalculate item similarity 
  • Ways to select the neighborhood 
  • Finding the right neighborhood 
  • Ways to calculate predicted ratings 
  • Prediction with item-based filtering 
  • Cold-start problems 
  • A few words on machine learning terms 
  • Collaborative filtering on the MovieGEEKs site 
  • What’s the difference between association rule recs and collaborative recs? 
  • Levers to fiddle with for collaborative filtering 
  • Pros and cons of collaborative filtering 

9. Evaluating and testing your recommender 

  • Business wants lift, cross-sales, up-sales, and conversions 
  • Why is it important to evaluate? 
  • How to interpret user behavior 
  • What to measure 
  • Before implementing the recommender… 
  • Types of evaluation 
  • Offline evaluation 
  • Offline experiments 
  • Implementing the experiment in MovieGEEKs 
  • Evaluating the test set 
  • Online evaluation 
  • Continuous testing with exploit/explore 

10. Content-based filtering 

  • Descriptive example 
  • Content-based filtering 
  • Content analyzer 
  • Extracting metadata from descriptions 
  • Finding important words with TF-IDF 
  • Topic modeling using the LDA 
  • Finding similar content 
  • Creating the user profile 
  • Content-based recommendations in MovieGEEKs 
  • Evaluation of the content-based recommender 
  • Pros and cons of content-based filtering 

11. Finding hidden genres with matrix factorization 

  • Sometimes it’s good to reduce the amount of data 
  • Example of what you want to solve 
  • A whiff of linear algebra 
  • Matrix 
  • What’s factorization? 
  • Constructing the factorization using SVD 
  • Adding a new user by folding in 
  • How to do recommendations with SVD 
  • Baseline predictors 
  • Temporal dynamic 
  • Constructing the factorization using Funk SVD 
  • Root Mean Squared Error 
  • Gradient descent 
  • Stochastic gradient descent 
  • And finally, to the factorization 
  • Adding biases  
  • How to start and when to stop 
  • Doing recommendations with Funk SVD 
  • User vector 
  • The items the user likes 
  • Funk SVD implementation in MovieGEEKs 
  • What to do with outliers 
  • Keeping the model up to date 
  • Faster implementation 
  • Explicit vs. implicit data 
  • Evaluation 
  • Levers to fiddle with for Funk SVD 

12. Taking the best of all algorithms: Implementing hybrid recommenders 

  • The confused world of hybrids 
  • The monolithic 
  • Mixing content-based features with behavioral data to improve collaborative filtering recommenders 
  • Mixed hybrid recommender 
  • The ensemble 
  • Switched ensemble recommender 
  • Weighted ensemble recommender 
  • Linear regression 
  • Feature-weighted linear stacking (FWLS) 
  • Meta features: Weights as functions 
  • The algorithm 
  • Implementation 

13. Ranking and learning to rank 

  • Learning to rank an example at Foursquare 
  • Re-ranking 
  • What’s learning to rank again? 
  • The three types of LTR algorithms 
  • Bayesian Personalized Ranking 
  • Ranking with BPR 
  • Math magic (advanced wizardry) 
  • The BPR algorithm 
  • BPR with matrix factorization 
  • Implementation of BPR 
  • Doing the recommendations 
  • Evaluation 
  • Levers to fiddle with for BPR 

14. Future of recommender systems 

  • Algorithms 
  • Context 
  • Human-computer interactions 
  • Choosing a good architecture 
  • What’s the future of recommender systems? 
  • User profiles 
  • context 
  • Algorithms 
  • Privacy 
  • Architecture 
  • Surprising recommendations 
View All Courses

    Course Inquiry

    Fill in the details below and we will get back to you as quickly as we can.

    Interested in any of these related courses?