Query Recommendation System

11/2022 - 01/2023

Project description

Hybrid recommendation system leveraging a linear combination of Expanded-Item-Item Collaborative Filtering and Compact Item-Item Collaborative Filtering based on query result cardinality.

Artifacts

Team and role

Team size: 2 people

• I was responsible for the implementation of the precursor of the generic Item-Item Collaborative Filtering algorithm and the Compact Item-Item Collaborative Filtering algorithm.
  
• I proposed and integrated the evaluation metrics used throughout the project.
  
• I implemented several other components, following other approaches, such as a Content-based algorithm, which did not satisfied our baseline requirements.
  
• I proposed and implemented the linear combination based on query result cardinality of the two best algorithms.
  
• Given my previous experience with data visualization, I created the visualizations of the results for the report, along with the contribution to the redaction of the report itself.

Tech stack