Skip to content

A recommender system for suggesting multi-feature paths based on user's interest.

License

Notifications You must be signed in to change notification settings

vcutrona/paths-rs

Repository files navigation

Multi-Feature Paths Composition and Recommendation

A recommender system for suggesting multi-feature paths based on user's interest.

Getting Started

Prerequisites

  • Python 2.7
  • PostgreSQL + PostGIS extension
  • Java

Dependencies and external data sources

All Python dependencies are listed in the requirements.txt file.

The map inference algorithm that we used is publicly available on Github.

All raw trajectories available under the docs directory were download from Wikiloc.

Running the application

A short demo is provided by the Demo class (under the main package).

You can also load our database dump in order to initialize your database. The dump was made with pgAdmin III (v. 1.22) and is available under the postgis_dump directory.

Algorithms Evaluation

Our experiments are available under the evaluation package. The Evaluator class exposes methods for running the experiments and computing the nDCG measure. The survey_results.csv file contains the users' answers that we collected through our survey. Since the annotation task is stochastic, you should also load our database dump in order to reproduce our experiment.

Authors

About

A recommender system for suggesting multi-feature paths based on user's interest.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages